RELEASE AND RECOVERY OF RHIZOBIUM FROM TROPICAL SOILS FOR ENUMERATION BY IMMUNOFLUORESCENCE

A THESIS SUBMITTED TO THE GRADUATE DIVISION OF THE UNIVERSITY OF HAWAII IN PARTIAL FULFILLMENT

OF THE REQUIREMENTS FOR THE DEGREE OF

MASTER OF SCIENCE

IN MICROBIOLOGY

AUGUST 1980

By

Mark T. Kingsley

Thesis Committee:

B. Ben Bohlool, Chairman L. R. Berger J. B. Hall

We certify that we have read this thesis and that in our opinion it is

satisfactory in scope and quality as a thesis for the degree of Master of

Science in Microbiology.

ACKNOWLEDGEMENTS

I am extremely grateful to Dr. B. Ben Bohlool for his guidance,

constructive criticisms, enthusiasm, and financial support throughout all

phases of this research.

I would like to thank Dr. L. R. Berger for his constructive

criticisms and help in reviewing the various sections of this thesis while

Dr. Bohlool was on leave.

In addition, I am extremely grateful to fellow graduate students

Renee Kosslak and Peter Alexander for their constructive criticisms of the

various stages of this thesis.

ABSTRACT

Immunofluorescence (IF) provides a direct method for in situ

autecological studies of microorganisms; it allows for the simultaneous

detection and identification of the desired organism in its natural habitat.

With the development of a quantitative membrane filter - immunofluorescence

technique, the range of applications of IF were extended to include

quantitative studies of microorganisms directly from soil.

The overall objective was to study the ecology of chickpea Rhizobium in

tropical soils. To accomplish this, the research described in this thesis was

concerned with: (1) determining the serological characteristics of 27 strains

of chickpea (Cicer arietinum L.) rhizobia, by immunofluorescence and

immunodiffusion, for use in ecological studies; (2) evaluation of the

quantitative membrane filter - immunofluorescence technique for studies of

Rhizobium in tropical soils; (3) the development of successful modifications of

the quantitative method to optimize release and recovery of Rhizobium from

tropical soils.

To employ the quantitative technique for the study of chickpea Rhizobium

in tropical soils, fluorescent antibodies (FA’s) were prepared from the somatic

antigens of the following strains: Nitragin strains 27A3, 27A8, 27A11, USDA

strain 3HOal; and NifTAL strains TAL-480, TAL-619, and TAL-620. Twenty-seven,

strains of chickpea rhizobia were screened with these seven FA’s; the

immunofluorescent reactions defined five groups. Group I, corresponding to

serogroup Nitragin 27A3, contained only the homologous strain. Group II,

serogroup Nitragin 27A8, Nitragin 27A11, TAL-619, and TAL-620, contained 15

cross-reacting strains. The four strains, Nitragin 27A8, Nitragin 27A11, TAL-

619, and TAL-620 were shown to have identical antigens by FA-cross adsorption,

and by immunodiffusion with whole cell antiserum. These four strains

constituted one serotype. Group III, serogroup TAL-480, contained two reactive

strains TAL-480 and TAL-622. Group IV, serogroup 3HOa9, was specific for the

homologous FA. Eight strains failed to react with any FA (Group V). No cross-

reactions were detected among 19 other strains of fast- and slow-growing

rhizobia.

FA and immunodiffusion were used to compare the antigens of two strains

of chickpea rhizobia obtained from both pure cultures and from nodules. The

immunofluorescent reactions of the nodules containing these strains paralleled

the reactions of their parent cultures. A difference was detected in the

quality of fluorescence between the nodule bacteria and their parent cultures.

The fluorescent outline of cells from culture was sharp and well defined, while

that of the nodule-bacteria was diffuse and thick. In immunodiffusion agar

gels, nodule antigens were freely diffusable while culture antigens required

heat-treatment.

The efficiency of the quantitative membrane filter technique for

recovering fast- and slow-growing rhizobia from tropical soils was evaluated

with eight soils, from three of the major soil orders (Oxisols, Inceptisols,

Vertisols). Recovery of added rhizobia from seven soils was less than or equal

to 13%. A recovery of 100% of the added cells was obtained with one

Inceptisol.

In a sand:soil (Oxisol) mixture, increasing the soil content from 0%

(i.e. 10 g sand) to 100% soil (10 g soil) caused a decrease in recovery of two

fast-growing strains of Rhizobium from 100% to less than 1%.

Modifications to the usual quantitative membrane filter--

immunofluorescence technique yielded consistently high and reproducible

recoveries of both fast- and slow-growing rhizobia from tropical

soils. The modified procedure involved suspending the soil by shaking with

glass beads on a wrist-action shaker. The diluent consisted of partially

hydrolyzed gelatin (0.1%)-0.1M (NH4)2HPO4. Growth of fast and slow growing

strains of Rhizobium in a sterile Hawaiian Oxisol was followed by plate counts,

the quantitative procedure and the modified quantitative procedure. Parallel

growth curves obtained with plate counts and the modified quantitative

procedure indicated close agreement, while counts with the original procedure

were 1000 times lower.

TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS ............................................... 3

ABSTRACT ....................................................... 4

LIST OF TABLES ................................................. 8

LIST OF ILLUSTRATIONS .......................................... 10

LIST OF ABBREVIATIONS AND SYMBOLS .............................. 11

CHAPTER 1. GENERAL INTRODUCTION ................................ 12

CHAPTER 2. LITERATURE REVIEW ................................... 14

CHAPTER 3. SEROLOGICAL ANALYSIS OF CHICKPEA RHIZOBIUM .......... 32

CHAPTER 4. PROBLEMS IN RECOVERING FAST-GROWING RHIZOBIA FROM TROPICAL SOILS FOR IMMUNOFLUORESCENT (IF) ENUMERATION ......................................... 52

CHAPTER 5. MODIFIED MEMBRANE FILTER - IMMUNOFLUORESCENCE FOR ENUMERATION OF RHIZOBIUM FROM TROPICAL SOILS ............................................... 71

APPENDICES ..................................................... 103

LITERATURE CITED ............................................... 109

LIST OF TABLES Table Page 1 Sources of cultures ....................................... 34 2 Immunofluorescence reactions of chickpea rhizobia ......... 38 3 Measure of similarity of the somatic antigens of

4 strains of chickpea Rhizobium from Serogroup II by FA/Cross-adsorption .................................... 40

4 Summary of antibiotic resistance patterns for

some chickpea Rhizobium strains used in this study ........ 50 5 Properties of soils used in Chapter 2 and 3 ............... 55 6 Recovery of TAL-620 from 8 different tropical

soils using SRP ........................................... 60 7 Recovery of TAL-620 from Wahiawa soil (Oxisol/

Hawaii): Evaluation of extractants for increasing recovery. I. Extracts yielding <1% recovery .............. 67

8 Recovery of TAL-620 from Wahiawa soil (Oxisol/

Hawaii): Evaluation of extractants for increasing recovery. II. Extractants yielding >1% recovery .............................................. 68

9 SRP - Effect of different strength Partially

Hydrolyzed Gelatin (PHG) solutions on increasing recovery of TAL-620 from Wahiawa soil .......... 80

10 SRP - Influence of pH of a 0.1% Partially

Hydrolyzed Gelatin (PHG) solution to recover TAL-620 from Wahiawa soil ................................. 81

11 SRP - Effect of different diluents to increase

recovery of TAL-620 from Wahiawa soil when mixed with Partially Hydrolyzed Gelatin (PHG) ............. 83

12 MSRP - Development of a modified soil release

procedure - effect of different Partially Hydrolyzed Gelatin (PHG) extractants on recovery of TAL-620 from Wahiawa soil .............................. 85

13 MSRP - Effect of the hydrated radius of four

monovalent cations upon recovery of TAL-620 from Wahiawa soil ......................................... 86

14 MSRP - Effect of shaking time on recovery of

TAL-620 from Wahiawa soil ................................. 87

Table Page 15 MSRP - Effect of gels from different sources:

Recovery of TAL-620 from Wahiawa soil ..................... 88 16 Procedure for the use of gelatin in the quantitative

procedure ................................................. 89 17 Growth of two strains of Rhizobium japonicum in

sterile Wahiawa soil, followed by Plate Counts (PC), Soil Release Procedure (SRP), and Modified Soil Release Procedure (MSRP) ............................. 100

18 Growth of USDA 110 in sterile Clarion soil,

followed by Plate Counts (PC), Soil Release Procedure (SRP), and Modified Soil Release Procedure (MSRP) .......................................... 102

APPENDIX TABLES 19a Yeast extract-mannitol medium (YEMS) ...................... 104 19b Defined agar medium ....................................... 105 20 Plant nutrient solution ................................... 106 21 Protocol for the release of soil bacteria for

enumeration by immunofluorescence microscopy (Soil Release Procedure SRP) .................................... 107

22 Formulation for phosphate buffered saline (PBS)

0.1M pH 7.2 ............................................... 108

LIST OF ILLUSTRATIONS Figure Page 1 Immunodiffusion analysis of strains of chickpea

Rhizobium from both culture and nodules ................... 42 2 TAL-620 broth culture, mid-exponential phase

cells stained with honologous FA .......................... 46 3 TAL-620 nodule smear, stained with FA prepared

against somatic antigens of TAL-620 from culture .......... 46 4 Nitragin 27A3 broth culture, mid-exponential phase

cells stained with homologous FA .......................... 48 5 Nitragin 27A3 nodule smear, stained with FA

prepared against somatic antigens of Nitragin 27A3 from broth culture ........................................ 48

6 Recovery of TAL-620 (Cicer) and Hawaii-5-0 (Lens)

from two Hawaiian Oxisols using SRP: soil titrations ...... 62 7 Recovery of TAL-620 from three different tropical

soils using SRP: soil titrations .......................... 64 8 Recovery of TAL-620 from two midwestern Mollisols

using SRP: soil titrations ................................ 66 9 Recovery of TAL-620 (Cicer) and Hawaii-5-0 (Lens)

from two Hawaiian Oxisols comparing SRP and MSRP: soil titrations ..................................... 91

10 Recovery of TAL-620 from three different tropical

soils comparing SRP and MSRP: soil titrations ............. 93 11 Recovery of TAL-620 from two midwestern Mollisols

comparing SRP and MSRP: soil titrations ................... 95 12 Growth of TAL-620 in a sterile Wahiawa Osixol

followed by PC, MSRP and SRP .............................. 98

LIST OF ABBREVIATIONS AND SYMBOLS

Å angstroms

CEC cation exchange capacity

FA fluorescent antibody

g gram

IF immunofluorescence

M molar

ml millilite

mm millimeter

MSRP Modified Soil Release Procedure

NaHMP Sodium Hexa-Meta Phosphate

O.M. organic matter

PBS Phosphate Buffered Saline

PC Plate Counts

PHG Partially Hydrolyzed Gelatin

SRP Soil Release Procedure

µl micrometer

µm micrometer

CHAPTER 1

GENERAL INTRODUCTION

The small size of bacteria dictates that they be viewed directly in nature

with the aid of a microscope. This is easily done in aquatic habitats.

Unfortunately the particulate nature of soil prevents easy viewing and

enumeration of microorganisms by conventional light microscopy. Cells may

attach to opaque soil particles and when stained by typical bacteriologic dyes

remain obstructed from view. Additionally, small pieces of organic and mineral

matter may be mistaken for bacteria.

To overcome some of these difficulties a number of specialized techniques

have been adopted to observe, study and enumerate bacteria microscopically in

soil. Several of these techniques take advantage of the smooth, artificial

surface of glass. Some examples are: the Perfil’ev capillary technique

(Perfil’ev and Gabe, 1969), the Cholodny buried slide (see Johnson and Curl,

1972) and the Breed slide (framer and Schmidt, 1964). Although useful, none of

these techniques offer the potential applications of fluorescent antibody (FA)

methodology. The application of immunofluorescence (IF) to the Rhizobium model

system (Schmidt et al., 1968) allowed investigators for the first time to

simultaneously observe and identify a microorganism of interest directly from

the soil amidst a plethora of other organisms.

Specific quantitative techniques necessary to measure biomass, growth rate

in soil, and growth responses to environmental variables are important to the

soil microbial ecologist. When a conventional Breed slide is stained with FA a

density of approximately 106 cells/gram of soil is necessary to encounter one

cell in ten microscope fields (100 X objective) (Bohlool, 1971; Schmidt, 1978).

In order to enumerate natural populations, usually less than 106 cells/gram of

soil, it becomes necessary to separate the bacteria from interfering soil

particles and concentrate them for enumeration.

In 1973 Bohlool and Schmidt (1973a) described a technique in which cells

recovered from soil on non-fluorescent membrane-filters, and stained with the

appropriate FA, were enumerated by immunofluorescence. Although applied in

several studies of rhizobia in soils and rhizospheres (Bohlool and Schmidt,

1973a; Reyes and Schmidt, 1979; Vidor and Miller, 1979) difficulties in the

efficiency of recovery of rhizobia were noted (Schmidt, 1974; Reyes and Schmidt,

1979; Vidor and Miller, 1979; Wollum and Miller, 1980). In addition, May (1978,

Personal Communication) and Kingsley and Bohlool (unpublished) obtained very

poor recoveries of lentil and chickpea Rhizobium respectively from a Hawaiian

Oxisol.

This research was concerned with: (1) the preparation of fluorescent

antibodies for, and determining the serological characteristics of strains of

Cicer rhizobia for use in ecological studies; (2) assessment of the sorptive

nature of several temperate and tropical soils for Rhizobium when assayed by the

quantitative membrane-filter technique (Bohlool and Schmidt, 1973a); and (3) the

development of successful modifications of the quantitative method so that

bacteria can be easily enumerated in tropical soils.

CHAPTER 2

LITERATURE REVIEW

Bacillus radicicola, the root-nodule bacteria of legumes, were first

isolated, described, and named by Beijerinck, the father of microbial ecology.

These organisms now constitute the genus Rhizobium--the name proposed by Frank

in 1889 (Fred et al., 1932). From Beijerinck’s report in 1888 to the present,

the Rhizobiaceae have been the object of intense investigation and are probably

among the most widely studied of the soil microorganisms.

Interest in these bacteria stems from the unique nitrogen-fixing

symbiotic association they have with their legume hosts. Legumes are among the

world’s most important crop plants, second only to grains (Advisory Committee

on Technology and Innovation 1979). Thus it seems only natural that the

symbiont be intensely studied. While the actual mechanisms of host specificity

remain elusive, questions concerning the life of these bacteria in the soil and

the soil properties which influence their growth, persistance, and success or

failure in nodulation can be answered. These answers can be readily applied to

increasing legume yields through enhanced symbioses.

I. Use of Serological Techniques in Studies of Rhizobium

Serological techniques have been in use for many years to investigate the

Rhizobiaceae. The discovery by Klimmer and Kruger (in Fred et al., 1932) that

bacteria isolated from different species of legumes could be distinguished

serologically, made serological methods extremely attractive for strain

identification. Stevens (1923) and later Wright (1925) found that different

strains isolated from the same species of plant, and therefore belonging to

the same inoculation group, were serologically unrelated. In fact, Hughes

and Vincent (1942) found strains isolated from different nodules on the same

plant which were serologically unique. The results of these early

investigations pointed to the great serological diversity now known to exist

in the Rhizobiaceae.

A. Agglutination

Agglutination was one of the first methods to be applied to serological

investigations of rhizobia. It is among the simplest of serological

techniques to use and it has been widely exploited in many taxonomic and

ecologic investigations. Bushnell and Sarles (1939) used the technique to

define three types of antigens on rhizobia. They reported on the antigenic

specificity between and within rhizobia from soybean, cowpea, and lupin

cross-inoculation groups. They found no correlation between the ability of

rhizobia from the three legumes to cross-inoculate and cross-agglutinate.

This important observation was recently restated by Vincent (1977): strains

which are related or apparently related serologically can be entirely

unrelated in other characteristics. Bushnell and Sarles (1939) also

confirmed the results of Stevens (1923) who found that due to the serological

diversity within a species of Rhizobium all strains cannot be identified by

the agglutination test.

Kleczkowski and Thornton (1944) used agglutination to study the

serological relatedness between and within pea and clover strains of

rhizobia. They tested six antisera (four clover, two pea) against 161

strains of R. trifolii, 29 R. leguminosarum, 5 each of R. meliloti and R.

lupini, and 13 non-Rhizobium soil isolates. Partial cross-reactions occurred

in the clover and pea groups which were removed after adsorption of antisera

with the cross-reacting antigens. No cross-reactions were detected outside of

the clover and pea groups; and none of the 13 soil isolates agglutinated. No

“group” antigen common to all the strains was found and attempts to link

effectiveness or ineffectiveness to any serological property failed.

Koontz and Faber (1961) used agglutination-adsorption (Edwards and Ewing,

1955) to characterize the somatic antigens of 25 strains of Rhizobium resulting

in six distinct serogroups. Antigenic similarities and physiological

characteristics could not be related.

Graham (1963), in a similar study to that of Kleczkowski and Thornton

(1944), prepared antisera against somatic antigens and whole cell/flagellar

antigens of 58 strains of root-nodule bacteria and 16 strains of agrobacteria.

He tested the antisera by tube agglutination against 113 strains of Rhizobium,

20 strains of Agrobacterium and 20 strains of other, possibly related bacteria.

On the basis of the agglutination reactions he categorized the rhizobia into

three serologically distinct groups. Cross-reactions were more common when

whole cell antisera were used than when agglutinations were run with somatic

antisera.

Tube agglutination techniques utilize antigens obtained from pure

cultures. Means et al. (1964) adapted the methods to type bacteria directly

from root-nodules. They observed that the agglutination reaction of pure

cultures and of root-nodule homogenates were identical for 15 of 17 strains

tested, and recommended that this technique be used for a quick classification

of nodules. This method was further modified into a micro-agglutination test

(Damirgi et al., 1967). In micro-agglutination a drop of dilute nodule

homogenate is mixed with a drop of dilute antiserum in a depression plate and

allowed to react. In this way the number of serologic tests of even small

nodules can be increased greatly.

B. Immunodiffusion

The technique of Ouchterlony double-diffusion has been widely used to study the

antigens of root-nodule bacteria for both taxonomic and ecologic purposes

(Dudman, 1964, 1971; Dudman and Brockwell, 1968; Gibbins, 1967; Humphrey and

Vincent, 1965, 1969, 1975; Vincent and Humphrey, 1968, 1970, 1973). Strains

which cross-react in agglutination tests because of minor similarities in their

surface, particulate antigens (such as flagella) do not necessarily share other

antigens as shown by immunodiffusion--a method which uses soluble antigens

diffusing through gels (Eisen, 1974; Dudman, 1977).

The gel diffusion method permits the enumeration and comparison of

antigens with minimal effort; but the confidence with which strains can be

identified will increase in proportion to the number of antigens detected

(Dudman, 1964, 1977; Eisen, 1974). Relationships between various antigens are

established by observing the nature of the interaction at the junction of

precipitin bands from the various wells, the number of precipitin bands being

equal to the minimal number of separately diffusable soluble antigens present

in the antigen well.

The somatic antigens of many Rhizobium strains diffuse slowly in the agar

gels; they yield either no precipitin bands or only weak bands close to the

antigen well since the location of bands is dependent upon the relative

concentrations of diffusable antigens and antibodies (Eisen, 1974). Heating

for various periods of time dissociates the poorly diffusable somatic antigen

molecules and makes them more soluble; thus it is one of the easiest methods of

antigen preparation for gel diffusion (Skrdleta, 1969; Dudman, 1971; Humphrey

and Vincent, 1975). Gibbins (1967) found ultrasonic disruption prevented

precipitin band formation; however, band formation was restored by heating the

sonicated antigen preparation. Sonication is a useful method to liberate

internal antigens which generally are not strain specific (Humphrey and

Vincent, 1965; Vincent and Humphrey, 1970).

Dudman (1964) was the first investigator to adapt gel diffusion to

studies of Rhizobium. His investigation of the extracellular soluble antigens

of two strains of R. meliloti indicated that the two strains shared all

extracellular antigens except for those strain-specific fast-diffusing

polysaccharides, which were useful for identification purposes. Since the two

strains did not cross-agglutinate he proposed that the strain-specific

polysaccharides dominated their surfaces.

Humphrey and Vincent (1965) used gel-diffusion to show that whole-cells

of R. trifolii strains grown on calcium-deficient medium yielded identical

immunodiffusion patterns with mechanically disintegrated calcium-adequate

bacteria. Earlier work by Vincent (1962) had shown that these strains required

calcium for normal growth. The identical immunodiffusion patterns indicated

that the walls of the untreated calcium-deficient bacteria were more fragile

and the cells underwent autolysis with the release of their internal antigens.

In a later publication Humphrey and Vincent (1969) indicated that the

somatic antigens of two strains of R. trifolii were strain specific. However,

the internal antigens obtained by mechanically disrupting the cells were

identical and could not be used to differentiate between strains.

Skrdleta (1969) utilizing gel-diffusion to study the serological

relatedness of strains of R. japonicum divided the 11 strains into two somatic

serogroups. The somatic antigens were more specific to differentiate between

individual strains than those of flagella. Dazzo and Hubbel (1975) in contrast

to the results obtained by Bushnell and Sarles (1939), Kleczkowski and Thornton

(1944), and Koontz and Faber (1961) reported a correlation between serological

properties and infectivity. They used immunodiffusion to analyze the antigenic

relatedness of three infective and three non-infective strains: additional

antigens were found in infective strains which were not found in non-infective

strains.

C. Immunofluorescence (IF)

One of the most sensitive of the serological techniques available to

study rhizobia is the fluorescent antibody (FA) technique. It allows for the

visualization and investigation of the antigens of individual cells with the

fluorescent microscope and requires only small quantities of both antigen and

antibody (Schmidt, 1973). In contrast both agglutination and immunodiffusion

require large amounts of antigen and antisera to give a visible reaction.

The FA technique originally developed by Coons et al. (1942) to visualize

pneumococcal antigens in mouse tissue, was successfully adapted to studies of

Rhizobium by Schmidt et al. (1968). For the first time individual cells could

be identified directly in culture, in nodules, and in soil be differentiated

from numerous other organisms. Using this technique, the classical approach of

autecology: the study of an individual organism in its natural environment,

could be applied to the study of rhizobia, or to any other soil microorganism

desired.

Although immunofluorescence (IF) can be used to rapidly type the

contents of root-nodules its most valuable feature is its potential ability to

identify specifically bacteria directly from soil (Bohlool and Schmidt, 1979).

None of the other serological techniques can be used to perform this function.

Fluorescent antibody (FA) has been used to identify strains of rhizobia

(Schmidt et al., 1968; Bohlool and Schmidt, 1970, 1973b; Jones and Russel,

1972, May, 1979), to identify the nodule-bacteria (Schmidt et al., 1968;

Trinick, 1969; Bohlool and Schmidt, 1973b; Jones and Russel, 1972; Lindemann et

al., 1974, May 1979), to detect doubly infected nodules (Lindemann et al.,

1974, May 1979), to study Rhizobium in soil (Schmidt et al., 1968; Bohlool and

Schmidt, 1970, 1973b; Vidor and Miller, 1979b and c), to study population

dynamics of R. japonicum in the rhizosphere (Reyes and Schmidt, 1979), and to

make quantitative studies of Rhizobium in soil (Bohlool and Schmidt, 1973a;

Schmidt, 1974; Reyes and Schmidt, 1979; Vidor and Miller, 1979b and c).

D. Enzyme-Linked Immunosorbant Assay (ELISA)

ELISA is the latest serological technique to be adapted for use in the

identification of rhizobia (Kishinevsky and Bar-Joseph, 1978; Berger et al.,

1979). The ELISA technique provides a colorimetric method for the

identification of bacteria. A strain specific antiserum is conjugated to an

enzyme such as alkaline phosphatase or peroxidase; and bacteria are coated with

the enzyme-labelled antibody. After a period of incubation and subsequent

washing, a chromogenic substrate is applied. The formation of an antigen-

antibody complex is detected visually or spectrophotometrically. The ELISA is

endowed with the specificity of antigen-antibody reactions and the sensitivity

of enzyme-catalyzed reactions (Berger et al., 1979). It requires very small

amounts of antiserum and no microscopic equipment is necessary. ELISA does not

possess the flexibility of immunofluorescence; like agglutination and

immunodiffusion, ELISA requires a purified antigen preparation, either from

culture or from a root-nodule.

E. Antigens of Rhizobia

In general the antigens of cultures and nodules remain stable and

antigenic stability is the major premise underlying the widely used serological

practices described previously for serotyping root-nodules.

However, some differences between antigens of the nodule forms of

Rhizobium and their parent cultures have been reported in the literature.

Means et al. (1964) used antisera against cultured cells to examine the

antigens of culture and nodule forms of 17 strains of R. japonicum. They found

no detectable difference between the two forms among 15 strains. However, with

one strain nodule-bacteria cross reacted with a wider range of antisera than

the parent culture. Both heated and unheated preparations reacted the same.

Dudman (1971) used immunodiffusion to examine the antigens of cultures

and nodule-bacteria of three strains of R. japonicum. Nodule-bacteria antigens

from one strain lacked the full array of antigens of the cultured cells, and

repeatedly formed spur reactions of partial identity with precipitin bands from

the parent culture. Using this technique, the classical approach of

autecology: the study of an individual organism in its natural environment,

could be applied to the study of rhizobia, or to any other soil microorganism

desired.

Although immunofluorescence (IF) can be used to rapidly type the contents

of root-nodules its most valuable feature is its potential ability to identify

specifically bacteria directly from soil (Bohlool and Schmidt, 1979). None of

the other serological techniques can be used to perform this function.

Fluorescent antibody (FA) has been used to identify strains of rhizobia

(Schmidt et al., 1968; Bohlool and Schmidt, 1970, 1973b; Jones and Russel,

1972, May, 1979), to identify the nodule-bacteria (Schmidt et al., 1968;

Trinick, 1969; Bohlool and Schmidt, 1973b; Jones and Russel, 1972; Lindemann et

al., 1974, May 1979), to detect doubly infected nodules (Lindemann et al.,

1974, May 1979), to study Rhizobium in soil (Schmidt et al., 1968; Bohlool and

Schmidt, 1970, 1973b; Vidor and Miller, 1979b and c), to study population

dynamics of R. japonicum in the rhizosphere (Reyes and Schmidt, 1979), and to

make quantitative studies of Rhizobium in soil (Bohlool and Schmidt, 1973a;

Schmidt, 1974; Reyes and Schmidt, 1979; Vidor and Miller, 1979b and c).

D. Enzyme-Linked Immunosorbant Assay (ELISA)

ELISA is the latest serological technique to be adapted for use in the

identification of rhizobia (Kishinevsky and Bar-Joseph, 1978; Berger et al.,

1979). The ELISA technique provides a colorimetric method for the

identification of bacteria. A strain specific antiserum is conjugated to an

enzyme such as alkaline phosphatase or peroxidase; and bacteria are coated with

the enzyme-labelled antibody. This indicated that the nodule-form and the

culture-form had slightly different antigenic structures. No antigenic

differences between culture and nodule forms were detected with the second

strain; and with the third strain the cultured cells occasionally yielded an

extra precipitin band.

Pankhurst (1979) could detect no differences in the antigens expressed by

cultures and nodules of a number of strains of Lotus rhizobia analyzed by

immunodiffusion. However, he noted that there may have been a structural

difference between the two forms. Whereas cultures had to be heat treated

before their antigens would diffuse through the gels, the nodule forms of

several strains were readily diffusable without heating. Heat treatment of the

nodule suspensions generally intensified the slow diffusing somatic precipitin

bands. He proposed that the readily diffusable nodule-bacteria antigens might

have been due to an alteration in structure, form, or amount (though not

antigenically different) of LPS as proposed by van Brussel et al. (1977) for

nodule-bacteria of R. leguminosarum.

No differences have been reported between nodule-bacteria and cultured

cell antigens when stained with FA’s prepared from the somatic antigens of

cultured cells (Schmidt et al., 1968; Bohlool and Schmidt, 1973b; Lindemann et

al., 1974, May 1979). However, no investigators have used immunofluorescence-

adsorption to examine differences between nodule and culture forms of Rhizobium

stained with fluorescent antibodies.

Serological markers also remain stable in soil. Brockwell et al. (1977)

found serological markers to be unchanged during a three year investigation of

rhizobia. Diatloff (1977) studied the stability of four rhizobial characters:

colony color, effectiveness, sensitivity to four antibiotics, and antigenic

stability. The antigenic and colony characteristics of the strains were not

changed during a residence of five to twelve years in the soil. On the other

hand, effectiveness and antibiotic sensitivity underwent slight modifications.

In summary the extensive use of serology to analyze rhizobia has revealed

that heat-stable somatic antigens are more strain specific than those of whole

cells (Koontz and Faber, 1961; Graham, 1963; Date and Decker, 1965; Means and

Johnson, 1968; Schmidt et al., 1968; Humphrey and Vincent, 1969; Skrdleta,

1969; Dudman, 1971, 1977). However, exceptions do exist, gel diffusion of the

somatic antigens from eight strains of R. meliloti (Humphrey and Vincent, 1975)

revealed seven of the eight strains had identical heat-stable somatic, as well

as heat-labile antigens. In this instance serology could not differentiate

between strains.

II. Intrinsic Antibiotic Resistance

Josey et al. (1979) used the variation in intrinsic resistance to low

levels of eight antibiotics as a characteristic (an “antibiotic fingerprint”)

to identify 26 strains of R. leguminosarum. The major advantage of this method

is that no alterations of the strain, which might interfere with its field

performance are required. The fingerprint technique is a useful supplement to

serological methods since, as with nutritional and biochemical tests, further

delineation of strains that serologically cross-react may be possible.

However, it is necessary to choose concentrations of antibiotics that will

yield maximum strain differentiation. Unlike serological techniques this

method requires isolation and culture steps which are time consuming.

III. Quantitative Techniques in Rhizobium Ecology

Only by studying the rhizobia directly in their natural habitat, rather

than by indirect plant-dilution infection techniques (Date and Vincent, 1962)

can the questions of survival, population densities, growth responses,

interactions with other organisms, nutritional substrates, host specificity,

and competition between strains for host sites be truly evaluated (Schmidt,

1978, 1979; Bohlool and Schmidt, 1979).

Although the need for a selective medium to enumerate Rhizobium has long

been recognized (early literature reviewed by Fred et al., 1932), the

physiological diversity of both the rhizobia and the resident soil

microorganisms makes the prospects for the development of a truly selective

medium specific for all rhizobia, unlikely (Schmidt, 1978). The work of Graham

(1969b) and Pattison and Skinner (1973) demonstrate the difficulties inherent

in making selective media for Rhizobium.

The “selective medium” most often used to detect the presence of rhizobia

in soil has been the plant itself. Enumeration by the plant-dilution infection

assay (or Most Probable Number, MPN) (Date and Vincent, 1962) has been the

basis for virtually all ecological studies dealing with the persistence of

Rhizobium in soil and their response to rhizosphere conditions (Schmidt, 1978).

Unfortunately there exists a high degree of statistical uncertainty in plant-

dilution infection/MPN methodology (Alexander, 1965). In addition the

existence of host Rhizobium incompatibilities resulting in nodulation failures

are well documented (Caldwell and Vest, 1968; Masterson and Sherwood, 1974;

Sherwood and Masterson, 1974). Nodulation by only certain strains of Rhizobium

might skew results when using plant-infection assays. This cannot happen with

direct microscopic enumeration of FA stained cells.

Before 1973 all serological techniques, used in studies of Rhizobium,

were qualitative. No means existed for the direct enumeration of rhizobia in

soils at natural population levels. Breed slide counts require high numbers of

rhizobia (106 cells/gram) to be practicable (Bohlool, 1971; Schmidt, 1978). To

work with natural populations it is necessary to remove the bacteria from

interfering soil particles and to concentrate them for enumeration. Bohlool

and Schmidt (1973a) described a soil release procedure and a quantitative

membrane-filter FA technique (SRP), in which cells, recovered from soil on non-

fluorescent membrane-filters and stained with the appropriate FA, were

enumerated microscopically. The development of the SRP technique was a

breakthrough in methodology for quantitative studies of microbial ecology in

soil.

A. Quantitative Membrane-Filter Technique

Briefly, the quantitative membrane-filter (SRP) technique consists of

dispersing by blending a soil sample in water. The soil suspension is then

transferred to a narrow container, flocculant is added and the soil colloids

allowed to settle. After settling, an aliquot of the supernatant is passed

through a membrane-filter, stained with FA and the FA reactive bacteria

enumerated (Bohlool and Schmidt, 1973a; Schmidt, 1974).

Since its development, the use of the quantitative technique in microbial

ecology has not often been reported in the literature. Bohlool and Schmidt

(1973a) followed the growth of a strain of R. japonicum (USDA 110) in

autoclaved soil (Clarion) comparing filter counts to plate counts. The filter

counts tended to underestimate at low cell numbers (approximately 30% of plate

counts) and overestimate the plate counts at high numbers (approximately 150%),

probably due to the accumulation of dead cells. The technique was used in a

field study to examine the population level of the same strain in the rhizo-

sphere of both inoculated and uninoculated (to examine the indigenous

population) soybean plants. The rhizosphere levels of the uninoculated plants

remained fairly constant (approximately 5 x 103/gram of soil) while the

rhizosphere populations of inoculated plants were both higher and more variable

(attributed to uneven inoculation).

Schmidt (1974) followed the growth of Nitrobacter winogradski in a

partially sterilized soil. He compared growth with nitrate formation; the two

parameters correlated well during the exponential phase of growth, however,

nitrate formation continued to increase when the growth of the population

apparently leveled-off.

Several reports exist in the literature describing problems in

implementing the membrane-filter quantitative technique. Reed and Dugan (1978)

used the quantitative method, with indirect immunofluoresence, to determine the

distribution of methane oxidizing bacteria in sediments from Cleveland harbor.

They recovered only 10% of the methane oxidizers in the sediments. With

several minor modifications Reyes and Schmidt (1979) obtained a 44% recovery,

as compared to plate counts, of a strain of R. japonicum (USDA 123) from a

sterilized Minnesota soil (Waukegan). The quantitative technique was

originally developed using a different soil:strain combination (Clarion:USDA

110). Reyes and Schmidt (1979) did not compare recoveries of USDA 110 in

sterilized Waukegan or USDA 123 in sterilized Clarion soil to determine if

their problem with recovery was methodological, or related to the soil, or the

strain. To aid recovery Vidor and Miller (1979a, b and c) substituted 1% CaCl2

as a flocculant.

Wollum and Miller (1980) modified the density gradient centrifugation

technique, used for studies of clay particles by Francis et al. (1972), to use

for quantitative studies of Rhizobium. They examined the recoveries of two

strains each, of the slow grower R. japonicum, and the fast grower R. phaseoli.

High recoveries were obtained when the soils contained 108 - 109 cells/gram.

Recoveries were better with temperate soils than with two South American

Oxisols. The authors had difficulty in clearing the Oxisols; they suggested

additional studies were needed.

B. Sorptive Interaction Between Microorganisms, Clay Particles and Soils

The predominance of the solid phase is one of the main characteristics

that distinguishes soil from other microbial habitats (Marshall, 1976; Stotzky

and Rem, 1966). In habitats which have wide liquid:solid ratios, such as the

oceans, much of the microbial activity is associated with solid, particulate

materials (ZoBell, 1943: Jannasch, 1967, 1970; Pearl, 1975). In fact,

attachment may be advantageous to microorganisms living in dilute environments.

Nutrients may concentrate at the interface because of differences in charge

between the solid surface and the surrounding solution, as well as to

differences in the hydrophobic or hydrophyllic nature of the surface (ZoBell,

1943; Stotzky, 1966a and b; Fletcher, 1979).

Of the solid phase components in soil, the smallest mineral particles are

the clays. It is due to their large surface area that clays exert the greatest

influence on soil microorganisms (Stotzky and Rem, 1966; Stotzky, 1966a and b).

Clay minerals are usually associated as aggregates or occur as coatings on

larger particles (Brady, 1974). Clays differ from other particles normally

present in natural microbial habitats in that they have an overall net negative

charge, but also have positive charges at the broken edges. These charges are

neutralized by ions from the surrounding solution, and by interactions with

adjacent minerals (Brady, 1974). In many tropical soils the minerals and clay

size particles are coated with an additional layer of oxides whose charges are

pH dependent (Sanchez, 1976). The charges on these coatings vary with local

soil conditions. Factors such as the type of clay, saturating ions, and the

organization of the clay minerals within the soil matrix may be more important

than the total amount of clay present in the soil habitat (Marshall, 1971;

Stotzky, 19743).

The degree of sorption between microorganisms and soil particles is

broadly related to the surface area and surface charge properties of the

particles (Daniels, 1972). The positive and negative charges on microbial

cells observed at different pH’s (Daniels, 1972; Lamanna and Mallette, 1965;

Marshall, 1967) are due to the degree of ionization of surface components.

Therefore the overall charge on the cell is determined by the isoelectric

points or pH’s of these constituents. Most organic materials of biological

origin have either no charge, or they are amphoteric and have a charge

dependent on pH (Lehninger, 1970; Metzler, 1977).

Fast growing strains of Rhizobium (R. trifolii, R. meliloti, R.

leguminosarum) exhibit a sharp increase in negative charge at high pH values

(pH 11) and a slightly positive charge at low pH’s (pH 2) when assayed by

electrophoretic mobility techniques (Marshall, 1967); this is generally true of

most bacteria (Lamanna and Mallette- 1965; Daniels, 1972). The slow growing

strains of Rhizobium (R. japonicum, R. lupini) possess either no charge at low

pH (pH 2) or a negative charge at all other pH’s. The rhizobia can therefore

be separated into two groups on the basis of their surface ionogenic character

(Marshall, 1967, 1968), as well as by differences in growth rate, G+C,

flagellation (Jordan and Allen, 1974), carbohydrate utilization (Fred et al.,

1932, Graham and Parker, 1964; Vincent, 1977), glucose-6-phosphate

dehydrogenase (Martinez-Drets and Arias, 1972; Martinez-Drets et al., 1977),

and internal antigens (Humphrey and Vincent, 1965; Vincent and Humphrey, 1970;

Vincent et al., 1973).

Marshall et al. (1971) proposed two major mechanisms for the sorption of

marine bacteria to glass surfaces: (1) a random attachment concerning both

motile and non-motile bacteria which involved polymeric bridging between the

cells and the solid substrate; (2) an instantaneous reversible mechanism

involving only motile bacteria which they believed was due to electrostatic

attractions and van der Waals forces (physico-chemical). Fletcher (1977) found

that the attachment of a marine pseudomonad to polystyrene petri dishes as

dependent on the number of cells present, the time allowed for attachment, the

growth phase of the culture (cell age and morphology), and temperature. Log

phase cultures had the greatest facility for attachment, followed by stationary

and death phase cultures, respectively. She found that the results could be

described by a model based on physical/chemical adsorption (Langmuir adsorption

isotherms). The model indicated that non-biological processes were playing a

major role in initial events of bacterial adhesion. Fletcher’s model is

similar to those describing molecular adsorption from solutions onto surfaces,

in which the process is controlled by the concentration of the solution (or

bacterial culture), time, and temperature. The relationship between solution

concentration and extent of adsorption is termed the adsorption isotherm. A

Langmuir Isotherm assumes: (1) adsorption is limited to a monomolecular layer;

(2) adsorption is localized so that adsorbed components are confined to

specific sites; and (3) the heat of adsorption is independent of surface

coverage (Fletcher, 1977).

Scheraga et al. (1979) found that bacteria added to autoclaved marine

sediments were adsorbed almost instantaneously. They plotted their data

according to Fletcher (1977) and obtained Langmuir-type adsorption isotherms.

This would indicate that the initial events in bacterial adsorption to

sediments are similar to those observed with polystyrene. Thus, it seems

likely that both biological and nonbiological forces are involved in sorptive

interactions between microorganisms and their environment.

Few reports exist describing actual sorptive forces in soil. The problem

is not necessarily one of methodology but one of deciding which fractions in

soil are most influential, for example clay, organic matter, clay:organic

matter complexes, oxide coatings, all fractions interacting simultaneously, or

particles of larger size coated with these substances. The factors which are

most influential in sorptive processes in one soil may have no bearing on those

operative in another soil.

Niepold et al. (1979) advanced a model system to describe various forces

which may influence the recovery of bacteria from soil. Bacteria may attach to

soil particles by capillary (Hattori, 1973; Hattori and Hattori, 1976),

electrostatic and adhesive forces (Marshall, 1971). Niepold et al. (1979)

reasoned that detachment of bacteria from soil particles might be influenced by

the chemical properties of the extraction fluid. They adsorbed the hydrogen

bacterium Alcaligenes eutrophus to three types of model materials with

differing capillary properties (“Mosy”, a porous material used for arrangements

of cut flowers, pore sizes 0.1-0.5 mm in diameter), electrostatic properties

(DEAE-cellulose), and adhesive properties, i.e. physico-chemical (125-200 mm

diameter glass beads). In these studies A. eutrophus was more strongly held by

capillary forces than by adhesive forces. Based on viable counts they found

recovery to be lowest when extraction fluids contained detergents such as Tween

80, Triton X-100, Sodium Dodecyl Sulfate (SDS), and sodium desoxycholate.

Recovery was also low when extraction fluids contained organic solvents such as

DMSO, or dioxane. The highest recoveries were obtained with Tris buffer (pH

7.5). The low recoveries with detergents and organic solvents were attributed

to the toxic effects these compounds had on the cells. Litchfield et al.

(1975) also found SDS to be toxic when used to quantitate bacteria in marine

sediments by viable counts.

Once extracted from soil, bacteria must be separated and recovered for

enumeration. Niepold et al. (1979) investigated various methods for the

separation of bacteria from model materials and soils. Filtration of soil

suspensions through filter paper, settling for eight hours, and slow speed

centrifugation (325 X g, 5 minutes) resulted in the lowest recoveries. In

contrast, letting the soil suspension settle for 15 minutes prior to dilution

for plate counts, resulted in greater recoveries. A flocculant to remove the

soil colloids from solution was never tried. They adsorbed eight strains of

hydrogen bacteria of differing size, motility, and slime formation to each of

three soils. The efficiency of extraction of the adsorbed bacteria was

strongly dependent on the bacterial strain used but was not significantly

dependent upon soil type.

Rubertschik et al. (1936) used plate counts to demonstrate the ability of

sediments to remove bacteria from suspension. The degree of sorption (e.g.

removal) varied with the sediment and the species of bacteria. Shaking the

mixtures for one minute did not increase desorption, i.e. the viable counts did

not increase.

Hornby and Ullstrop (1965) found that for dilution plating from soil

suspensions, agitation by blending or rapid stirring gave better sampling

precision than shaking with a reciprocating shaker. They also found that a

viscous solution, such as 1% carboxy-methyl cellulose or 0.2% agar gave better

reproducibility than soil suspended in water. Sodium hexa-meta-phosphate (Na-

HMP) is a powerful soil deflocculant (Michaels, 1958; Lahav, 1962). The many

negative charges of phosphate ions combine with the broken edges of clay

particles (positively charged) and neutralize their charges, thus soil

particles are dispersed by mutual repulsion. Gamble et al. (1952) briefly

blended (45 seconds) 10 grams of soil in 100 mls of 5% NaHMP, added 400 mls of

water to produce a total dilution of 1:50, and then blended for another two

minutes before plating. They felt this procedure would release the bacteria

for plate counts. However, they never compared their method with any others.

Both Balkwill et al. (1975) and Faegri et al. (1977) used a combination

of sequential homogenizing procedures with blendors, and sedimentations by

centrifugation to extract and concentrate bacteria from soils. Faegri et al.

(1977) enumerated bacteria extracted from Norwegian soils by direct counts of

acridine orange (AO) stained cells (AO is a non-specific fluorescent dye)

recovered on black membrane-filters. Depending upon the soil type almost equal

numbers of cells were obtained in the first two extractions. Fewer cells were

recovered following a third extraction.

From the foregoing discussion, it appears that there are many different

opinions as to what method is best to recover and enumerate bacteria from soil.

Depending upon the type of counting procedure to be employed, the type(s) of

microorganism(s) to be enumerated and the type of soil which the bacteria

inhabit, it is problematical whether only one or more of the described

procedures could work. Except for Wollum and Miller (1980) all the

investigations cited above have dealt mainly with temperate soils. There is a

need, therefore, to determine which methods will function in tropical soils for

the quantitative recovery of Rhizobium using the FA-membrane filter procedure.

CHAPTER 3

SEROLOGICAL ANALYSIS OF CHICKPEA RHIZOBIUM

Introduction

Serology has been used extensively in many taxonomic and ecological

investigations of the Rhizobiaceae. The fluorescent antibody (FA) method is the

most flexible of the serological techniques since it permits the investigation of

Rhizobium in its several habitats (Schmidt, 1973; Dudman, 1977; Bohlool and

Schmidt, 1979). The gel diffusion technique has also been extensively applied to

taxonomic (Dudman, 1964, 1971; Gibbins, 1967; Humphrey and Vincent, 1965, 1969,

1975; Vincent and Humphrey, 1968, 1970, 1973) and ecological (Dudman and

Brockwell, 1968; Dudman, 1977; Vincent, 1977) investigations of rhizobia. These

serological investigations have emphasized the marked somatic heterogeneity of the

Rhizobiaceae. When groups of microorganisms which have not been investigated

previously are to be studied serologically, it is necessary to determine their

degree of serological relatedness; ecological applications of fluorescent antibody

(FA) methodology require an especially high degree of serological specificity

(Schmidt, 1973).

Cicer arietinum L. (chickpea, garbonzo) is an important pulse crop; and it

is the third most widely-grown grain legume in the world (van der Maesen, 1972).

It is grown extensively in the Middle East; and in many regions of India it is the

most important legume grown (Medhane and Patil, 1974; van der Maesen, 1972).

Neither Cicer rhizobia nor the host have received much attention in their

literature.

This report describes a serological investigation of 27 strains of Cicer

rhizobia with seven anti-chickpea Rhizobium FA’s. Immunofluorescence adsorption

(Belly et al., 1973) and immunodiffusion were used to more fully evaluate the

relationships of four of the strains used to produce FA’s. In addition, the

antigenic properties of two strains and their corresponding nodular forms were

investigated using both immunofluorescence and immunodiffusion. As a further

means of differentiation and identification, 15 strains were tested for their

resistance to low levels of nine antibiotics (Josey et al., 1979).

Materials and Methods

I. Source and Maintenance of Cultures

Table 1 lists the strains, their origin and source, and a detailed pedigree

of the Cicer rhizobia used in this study. All strains were maintained on a yeast

extract-mannitol medium (YEMS) (Bohlool and Schmidt, 1970) (see Appendix Table 19a

for composition). For experimental use strains were grown in YEMS broth. All

media were sterilized by autoclaving at 121oC for 20 minutes.

II. Production of Root-Nodules

Seeds of Cicer arietinum L. variety JG-62 (provided by Dr. P. J. Dart

ICRISAT, Hyderbad, India) were surface sterilized with 4% calcium hypochlorite for

15 minutes, rinsed six times in sterile water, and germinated aseptically in petri

dishes containing 1% water agar. The seedlings were planted in modified Leonard

jars (Leonard, 1943), which contained sterile vermiculite and a nitrogen-free

nutrient solution (Broughton and Dilworth, 1971) and had the following

composition: for each 10 liters of complete culture solution 5.0 m. each of

solutions 1 to 4 (see Appendix Table 20), was added to 5.0 liters of water and

diluted to 10 liters. The pH was adjusted to 7.0.

Each of three seedlings in a jar received 1 ml of a cell suspension

containing 1 x 106 rhizobia/ml; two jars were left uninoculated. A 2 cm layer of

sterile perlite was added to the surface of all jars. Plants were grown under

controlled conditions in a growth chamber (Model M-31 Environmental Growth

Chambers, Chagrin Falls, Ohio) with a 14 hour day and a day/night temperature of

290C/240C. The nodules were harvested after three weeks and used immediately for

immunofluorescence and immunodiffusion. Nodules for immunodiffusion were

crushed in filtered saline containing thimerosal (1:10,000) and were either left

untreated or received a heat treatment similar to culture antigens.

III. Serological Procedures

A. Immunofluorescence

Fluorescent antibodies were prepared against the somatic components of TAL-

480, USDA 3HOa9, TAL-619 and TAL-620, and Nitragin strains 27A3, 27A8, 27A11.

Preparation of antisera and conjugation procedures were done according to Schmidt

et al. (1968) except cultures were grown in YEMS broth for three days instead of

seven days. The rabbits which were first injected with somatic cell antigens

were later injected with whole cells to develop antisera for immunodiffusion. No

whole cell serum was prepared for strain USDA 3HOa9.

Smears from pure cultures and nodules were stained by the method of Schmidt

et al. (1968) using gelatin-rhodamine isothiocyanate conjugate to control non-

specific staining (Bohlool and Schmidt, 1968). Stained smears were observed with

a Zeiss universal microscope equipped for epifluorescence and transmitted dark

field. Incident illumination was from an HBO-200 (OSRAM) light source with a

fluorescein isothiocyanate (FITC) filter. Transmitted dark field was from a l2v

quartz-halogen lamp, using a Zeiss Ultra-condenser.

Cross-reactions of FA-stained smears were quantitated by subjective

assessment of the degree of fluorescence, from 0 to 4+. Immunofluorescence

adsorption (Belly et al., 1973) was used for finer antigenic analysis and to

determine serologic relatedness. The somatic antigens used for adsorption were

prepared from three-day old shake-flask cultures (Schmidt et al., 1968),

distributed to three tubes and pelleted by centrifugation. The pellet from one

tube was resuspended in 6 ml of undiluted FA and incubated at room temperature.

Thimerosal was added as a preservative, final concentration 1:10,000). The

fluorescent agglutinate was removed after four hours, the supernatant transferred

to a fresh pellet, and incubated as before. The adsorption cycle was carried out

three times, with the final incubation proceeding overnight at 40C. The adsorbed

FA was used at a dilution of 1:4.

B. Immunodiffusion

Antisera for immunodiffusion were prepared from whole cell antigens grown on

solid, defined medium (Vincent, 1970) (see Appendix Table 19b) for composition.

The same rabbits used for production of somatic antisera for FA were injected

intramuscularly with one ml (0.5 ml each hip) of an equal mixture of antigen and

Freund’s complete adjuvant (Difco). After three weeks two mls of the same

culture was injected intravenously without adjuvant. The rabbits were bled one

week later.

Immunodiffusion was performed by the method of Dudman (1971). Cell

suspensions used in gel diffusion experiments contained approximately 1010 cells

per milliliter, equivalent to approximately 25 mg dry weight cell mass per ml.

Crushed nodule suspensions were prepared in saline (Dudman, 1971) to which

thimerosal was added (1:10,000) as a preservative. For the study of somatic

antigens, cell suspensions were heated in tightly capped tubes (Cryotubes,

Vanguard International, Neptune, N.J.) for two hours in a water bath at 1000C

(Skrdleta, 1969). The gels were incubated at room temperature (about 230C) for

one week. Results were recorded photographically.

IV. Intrinisic Antibiotic Resistance

The method used was similar to that described by Josey et al. (1979). The

use of a multiple inoculator allowed for the simultaneous inoculation of up to 27

cultures per petri plate. In practice, 15 cultures were replicated four times

per antibiotic concentration used. Fresh solutions of antibiotics were added to

cooled (480 C), melted YEMS medium (Bohlool and Schmidt, 1970, see Appendix Table

19a) to give final concentrations (mg/ml) of: chloramphenicol .012, .025;

streptomycin sulfate .0025, .010; tetracycline hydrochloride .004 (antibiotics

obtained from Calbiochem); kanamycin sulfate .010; naladixic acid .010; neomycin

sulfate .0025; polymyxin B sulfate .020; rifampin .006, .010; vancomycin .0015,

.005 (antibiotics obtained from Sigma). In the tables the names of the

antibiotics are abbreviated to the first three letters. Antibiotic stock

solutions were prepared in sterile distilled water at a concentration of 10

mg/ml, except for chloramphenicol (10 mg/ml in 95% ethanol), naladixic acid (10

mg/ml in 1N NaOH, and rifampin (10 mg/ml in methanol). Petri plates contained 25

ml of medium.

Results and Discussion

All Rhizobium species or inoculation groups examined so far have been found

to contain strains that are serologically distinct, as well as strains that share

cross-reactive antigens, i.e. no Rhizobium species is serologically homogeneous

(Dudman, 1977). The chickpea rhizobia are no exception. Twenty-seven strains of

Cicer rhizobia were tested with seven anti-chickpea Rhizobium FA’s. The four

serogroups defined by the resulting immunofluorescent reactions are listed in

Table 2. Those strains that failed to react with any FA were designated

Serogroup V. No cross-reactions were detected with 19 other non-Cicer rhizobia.

Serogroups I and IV were highly specific for their homologous antigens, Nitragin

27A3 and USDA 3HOa9, respectively. Serogroup II contained 15 strains which

cross-reacted at maximum fluorescence (4+) with FA’s TAL-619, TAL-620, and

Nitragin 27A8 and 27A11.

The large number of cross-reactions between strains of Serogroup II were

quite unexpected. By tracing the histories of a number of these strains (Table

1) it was determined that several of them originated from one culture. The

problem arose from the use of multiple collection numbers for the same culture (a

phenomenon which should be of concern to those working with Rhizobium, especially

curators of culture collections). For example, six collection numbers have been

applied to USDA 3HOa1 (which itself originated in India). It is the same as ATCC

11444, Rothamsted 3827, ICRISAT 3827, CC-1189, TAL-385, and TAL-619. Several of

the other strains in this serogroup may be related, but an incomplete pedigree

precludes discerning such a relationship. Therefore six of the cross-reactions

were only apparent.

Nitragin strains 27A8 and 27A11 (isolated in Mexico in 1971) and NifTAL

strains TAL-619 (source = USDA 3Hoa1, ex-India M.S. Raju, origin ?), and TAL-620

(ex-Israel, origin ?) (see Table 1) although from different sources and isolated

in different parts of the world are serologically identical by immunofluorescence

(Table 2), immunofluorescence-adsorption (Table 3) and immunodiffusion (Figure

1A, B). These strains, therefore, constitute one serotype. Normally serological

cross-reactions are of a lower titer than homologous reactions (Eisen, 1974).

Since these four strains had identical cross-reactive titers, and since they

produced identical intrinsic antibiotic sensitivity patterns (Table 4) they are

all probably derived from the same strain. This effective nodulator may have

been disseminated around the world by various researchers. Further biochemical

tests, or perhaps ribosomal protein mapping could elucidate the relationships of

these strains. A more complete history of a number of these strains would be the

easiest way of determining potential relationships between those that cross-

react.

The antigenic relationship between the cultured cells and bacteroid forms

of rhizobia is a subject of great interest. Means et al. (1964), and Dudman

(1971) working with strains of R. japonicum, and Pankhurst (1979) working with

strains of Lotus rhizobia have examined this relationship. Means et al. (1964)

examined the cultured and nodule forms of seventeen strains of R. japonicum by

agglutination and found no detectable difference between the two forms in fifteen

strains. However, in one strain the nodule-bacteria failed to cross-react with

the homologous antiserum. In another strain, the nodule form cross-reacted with

a wider range of anitgera than the cultured cells. Dudman (1971) compared the

antigens of nodule-bacteria and cultured forms of three strains of R. japonicum

Figure 1. A - Identical immunodiffusion patterns of four cross-

reacting strains of chickpea rhizobia from Sero- group II, used for antiserum production.

Center well = Nitragin 27A8 antiserum, Well 1 = Nitragin 27A8, Well 2 = Nitragin 27A11, Well 3 = TAL-619, Well 4 = TAL-620. (All antigens heat treated 2 hours, 1000C).

B - Antigen wells same as for A. Center well = TAL-620

antiserum.

C - Comparison of Nitragin 27A3 antigens from culture and from nodules.

Center well = Nitragin 27A3 antiserum, Wells 1 & 4 =

antigens from solid defined medium (heat-treated 2 hours, 1000C), Well 2 = antigens from untreated broth cultures, Well 3 = broth culture (heat treated 2 hours, 1000C), Well 5 = antigens from crushed nodules (untreated), Well 6 = antigens from crushed nodules (heat treated 2 hours, 1000C).

D - Comparison of TAL-620 antigens from culture and from nodules.

Center well = TAL-620 antiserum, Wells 1 & 4 = antigens from solid defined medium (heat treated 2 hours, 1000C), Well 2 = antigens from crushed nodules (heat treated 2 hours, 1000C), Well 3 =

antigens from crushed nodules (untreated), Well 5 = antigens from broth culture (untreated), Well 6 = antigens from broth culture (heat treated 2 hours

1000c).

using gel diffusion. The nodule form of one of the strains lacked the full array

of antigens associated with the cultured form. With a second strain no

antigenic differences were detected between the two, while the nodule form of a

third strain occasionally yielded an extra precipitin band. Pankhurst (1979),

investigating Lotus rhizobia, found no differences in the total array of antigens

expressed by the two forms. However, in contrast to the cultured cell forms of

these strains, he found that the nodule form of several strains required no pre-

treatment to give strong somatic precipitin bands.

The immunofluorescence reactions of nodules containing Nitragin 27A3 or

TAL-620 paralleled the reactions of their parent cultures. Although no

differences were apparent in the amount of fluorescence between nodules and

cultures (i.e. 4+) a difference was detected in the quality of fluorescence.

Whereas the fluorescent outline of cells from culture was sharp and well defined,

the fluorescent surface of the nodule-bacteria appeared diffuse and thick,

perhaps indicating a difference in cell wall structure between the two forms.

Nodule bacteria have long been known to be pleomorphic (Fred et al., 1932). The

shapes of the Cicer nodule-bacteria were different from their parent cultures.

The parent culture of TAL-620 contained rods (dimensions 3 x 1 µm, ± 0.7 x 0.1

µm) (see Figure 2) while the nodule bacteria were spherical (diameter = 1.7 ± 0.5

µm) (see Figure 3). The cultured form of Nitragin 27A3 had dimensions 2.6 x 1 µm

± 0.7 x 0.1 µm (see Figure 4) while the nodule bacteria tended to be thickened

rods slightly larger than the parent culture 3 x 1.3 ± 0.5 x 0.2 µm (see Figure

5).

Serological differences between Rhizobium in culture and in nodules have

been shown to occur (Means et al., 1964; Dudman, 1971; Pankhurst, 1979), and van

Brussel et al. (1977) showed that the cell walls of nodule-bacteria of R.

leguminosarum have different amounts of LPS than cell walls obtained from

cultures grown on regular media. The two strains of chickpea Rhizobium from

nodule-bacteria and from culture exhibited differences in quality of

immunofluorescence. For these reasons a study, using immunodiffusion, was

undertaken to investigate the antigens of these two strains of chickpea Rhizobium

from culture and from nodules.

If Nitragin 27A3 antigens (Serogroup I), grown on solid defined medium

(Appendix Table 19a), were left untreated no precipitin bands developed.

However, untreated TAL-620 antigens (Serogroup II) produced one precipitin band

close to the antigen well. After Nitragin 27A3 antigens were heat treated for

two hours, three precipitin bands developed. Similarly, heat-treated TAL-620

antigens produced three precipitin bands after the two hour heat-treatment; the

outermost band produced reactions of identity with the single band from the

unheated antigen wells. For both Serogroups I and II no differences were noted

in precipitin patterns of heat-treated antigens obtained from YEMS broth

(Appendix Table 19a), solid defined medium (Appendix Table 19b), or a heated

saline suspension of crushed nodules (see Figure 1C,D). Unheated broth cultures

produced variable patterns depending upon the age of the culture. Young broth

cultures (unheated) produced patterns similar to unheated cultures grown on solid

defined media (not shown in Figures). Older cultures, especially TAL-620,

yielded one band close to the antigen well and a unique heat labile band.

Unheated TAL-620 nodule suspensions also produced a unique pattern (Figure 1D).

In addition to the three bands found in both heated nodules and cultures unheated

nodule suspensions released a unique heat labile band. Since somatic antigens

are known to be heat stable (Koontz and Faber, 1961; Date and Decker, 1965;

Skrdleta, 1969; Dudman, 1977), this additional band may have been an internal,

heat labile antigen released from lysed nodule bacteria. The TAL-620 broth

culture may have been too old and many cells may have lysed. Unheated antigens

from defined agar media formed one band close to the antigen well and no heat

labile, fast diffusing antigens were observed.

The identification of strains of Rhizobium in soil or nodules using

serology is only reliable when no cross-reactions are known to occur. Where

Figure 2. TAL-620 broth culture, mid-exponential phase. Cells stained with homologous FA. Note typical rod shape. (scale = 4 µm)

Figure 3. TAL-620 nodule smear, stained with FA prepared against somatic antigens of TAL-620 from culture. Note round shape of the nodule-bacteria. (scale = 4 µm)

Figure 4. Nitragin 27A3 broth culture, mid-exponential phase. Cells stained with homologous FA. Note typical rod shape. (scale = 4 µm)

Figure 5. Nitragin 27A3 nodule smear, stained with FA prepared

against somatic antigens of Nitragin 27A3 from broth culture. Note pleomorphic shape of nodule-bacteria. (scale = 4 µm)

strains do cross-react another method of identification is required. The use of

intrinsic antibiotic resistance (Josey et al., 1973) is a simple and useful

technique to determine the similarity of strains. The results of such tests

with Cicer rhizobia (Table 4) agree with observations made by Vincent (1977)

that strains identical serologically can be different in other properties.

Serology should not be the sole criterion to determine strain

interrelationships. Within Serogroups II and III, where strains appeared

similar or identical by immunofluorescence, different antibiotic fingerprints

were obtained which indicated physiological dissimilarities between strains

(Table 4). In some cases several strains produced identical resistance patterns

and were grouped together.

Selection of the optimum antibiotic concentration to permit maximum

differentiation between members of a species of Rhizobium is important. Josey

et al. (1979) suggest that concentrations of antibiotics useful to differentiate

between strains of R. leguminosarum were not as useful to differentiate between

strains of R. phaseoli; to achieve this, higher antibiotic concentrations may

have been required. The simultaneous testing of several antibiotic

concentrations should permit selection of the proper concentrations. The

concentrations of several antibiotics in this study were not optimum, as can be

judged from the number of completely resistant and sensitive strains tested.

For resistant strains higher concentrations of antibiotics would have been

necessary to observe sensitivity. Where all strains were sensitive to a given

concentration, a lower level of antibiotic should have been used.

Summary and Conclusions

The serological reactions of Cicer rhizobia were species specific and

generally strain specific. The occurrence of a large number of apparent cross-

reactions was due to the same strain (USDA 3Hoal) having multiple accession

numbers from various Rhizobium collections.

The antigens from cells in nodules of two strains, Nitragin 27A3 and

TAL-620, had different structures than their corresponding cultures. Nodule

antigens were freely diffusable in immunodiffusion gels while antigens from

culture required a preparative heat treatment. Pankhurst (1979) found Lotus

nodule-bacteria to behave similarly.

A number of strains were screened for resistance to low levels of nine

antibiotics. Strains which cross-reacted serologically in some cases

produced identical patterns while in others they produced unique patterns.

This would indicate a physiological heterogeneity exists within these

serologically cross-reactive strains.

CHAPTER 4

PROBLEMS IN RECOVERING FAST-GROWING RHIZOBIA FROM TROPICAL SOILS FOR IMMUNOFLUORESCENT (IF) ENUMERATION

Introduction

Microbial ecologists require accurate techniques to quantify soil

microorganisms. Such techniques help provide estimates of biomass, growth rates

in soil and growth responses to environmental variables. Although plate counts

are one of the simplest and perhaps most highly exploited of the quantitative

techniques, they are particularly inadequate for soil microorganisms. It is

impossible even to approximate natural environmental parameters in culture media

(Brock, 1971; Schmidt, 1978). In addition there exists a discrepancy between

numbers of microorganisms indicated by colony counts and those obtained by direct

microscopic counts. Direct counts of soil bacteria usually result in total

numbers at least three times greater than those obtained by plate counts (Stotzky,

1972; Faegri, et al., 1977). In one case up to 100 times more bacteria and six

times more fungi were estimated by a direct microscopic examination of soil

(Skinner, et al., 1952). Similar discrepancies have been observed for the natural

bacterial flora in seawater (Jannasch and Jones, 1959).

Attempts have been made to produce media for the selective enumeration of

Rhizobium directly from soil (Graham, 1969b; Pattison and Skinner, 1973; Barber,

1979). However, the biochemical diversity of these agriculturally important

bacteria prevents the formulation of a medium truly selective for all strains

(Schmidt, 1978). Because of the lack of a selective medium or other adequate

procedures, the plant-infection technique has been the basis for virtually all

ecological studies of the root-nodule bacteria (Schmidt, 1978). This method

relies on the legume to act as a selective agent for Rhizobium. Plant-infection

is both cumbersome, since a large number of seedlings is required, and time

consuming, since one has to wait for the seedlings to grow and nodulate.

Bohlool and Schmidt (1973a) developed a soil release procedure (SRP) for the

rapid quantification of specific strains of Rhizobium directly from soil. This

procedure takes advantage of the specificity of immunofluorescence to detect and

identify the desired organism from others present, and the ability of membrane

filters to concentrate bacteria which are released into solution from the soil

matrix. When applied to studies of slow-growing Rhizobium in several soils of the

midwestern United States a 100% efficient release of the cells was not always

attained (Vidor and Miller, 1979a; Reyes and Schmidt, 1979). Inefficient release

has also been documented with other bacteria such as Azotobacter (Schmidt, 1974),

and Methane oxidizers (Reed and Dugan, 1978). In addition Kingsley and Bohlool

(Proc. 7th No. American Rhizobium Conf., p. 8, 1979) found this technique was not

satisfactory to recover strains of fast-growing Rhizobium from tropical soils for

IF enumeration. The problem of inefficient release must be solved since it

greatly limits the ability to study the ecology of free-living Rhizobium in

tropical soils.

Materials and Methods

Source and Maintenance of Cultures

Two strains of fast-growing Rhizobium were used in this series of

experiments: TAL-620 (Rhizobium for Cicer, see Chapter 1, Table 1), and Hawaii-5-0

(Rhizobium leguminosarum, for lentils, May, 1979). Both strains were maintained

on a modified YEMS medium (see Appendix Table 19a) and when required were grown in

broth of the same composition. Exponential phase broth cultures were enumerated

by counting with a Petroff-Hauser chamber. The cultures were diluted in filtered

(0.45 µm) saline to the desired inoculum size. For experiments one or two ml of

the inoculum was added to moistened soil.

Chemical Reagents

Reagent grade chemicals were purchased from: J. T. Baker, Phillipsburg, N.

J.; Mallinckrodt, St. Louis Missouri; Matheson, Coleman & Bell, Norwood, Ohio.

Antifoam C Emulsion, Thimerosal, and Tween 80 were obtained from Sigma Chemical

Company, St. Louis, Missouri. Disodium-EDTA was obtained from Eastman Organic

Chemicals, Rochester, N. Y. Nonidet P40 (BDH), a non-ionic detergent, was

purchased from Gallard-Schlesinger Chemical Mfg. Corp., Carle Place, N.Y. Peptone

(Bacto-Peptone) and gelatin (Bacto-Gelatin) were purchased from Difco, Detroit,

Michigan.

Soil Samples

Table 5 lists the ten soils used, their pH, cation exchange capacity (CEC),

and percent organic matter (% O.M.) (for place of origin see Table 6, except for

Clarion and Hubbard--soils from Minnesota). All soils were air-dried, sieved

through a #25 mesh (710 µm) sieve, and ten gram portions were dispensed to 25 x

200 mm screw-cap tubes. Prior to inoculation the soils were moistened with

distilled water such that the volume of liquid in the inoculum would give a final

soil moisture content of approximately 60% of the water holding capacity (Bohlool,

1971.).

Preparation of Fluorescent Antibodies (FA, FA Staining, Epifluorescence Enumeration

Fluorescent antibodies were prepared by the method of Schmidt et al. (1968).

The basic protocol for the release of strains inoculated into soil, and their

subsequent concentration and immunofluorescent enumeration on membrane-filters,

followed the procedures outlined by Bohlool and Schmidt (1973a) and Schmidt (1974)

(see Appendix Table 21). Several modifications to these procedures should be

noted. Polycarbonate membrane-filters (Huclepore) stained with Irgalan Black

(Hobbie et al., 1977) were substituted for india-inkstained Millipore filters.

The gelatin-rhodamine isothiocyanate conjugate (Bohlool and Schmidt, 1968) was let

dry completely on each filter prior to FA staining (May, 1979). Gelatin

coated filters could be stored dry, either in a desiccator, or left in a

drying-oven (600C) for long periods of time prior to staining. This greatly

simplified the processing of large numbers of samples. All FA enumerations

were made with a Zeiss universal microscope. Incident illumination was from

an HBO-200 (OSRAM) light source, and a Zeiss Fluorescein Isothiocyanate (FITC)

filter.

Recovery of TAL-620 from Eight Tropical Soils

To determine the sorptive nature of various tropical soils for a strain

of chickpea Rhizobium, eight soils, representative of three soil orders common

to the tropics, were chosen: Wahiawa (Oxisol); Molokai (Oxisol); Lualualei

(Vertisol); PLP, Burabod, LPHS, Makiki, and Waimea (all Inceptisols). An

exponentially-growing culture of TAL-620, enumerated with a Petroff-Hauser

chamber, was adjusted to 2 x 107 cells/ml with saline (= Inoculum). The

actual inoculum size was verified by FA-membrane-filter counts of the

inoculum. Each tube of ten grams of non-sterile soil received two ml of

inoculum

(= 4 x 106 cells/gram soil). Two tubes containing ten grams of silica sand

served as inoculated non-soil controls. After incubation for two hours at

room temperature (240C) the recovery of TAL-620 was assayed using the

procedure of Bohlool and Schmidt (1973a) with the modifications previously

described (see Appendix Table 21).

Soil Titrations

This series of experiments was designed to illustrate the sorptive

capacities of soils, both tropical (5 soils) and temperate (2 soils), for fast

growing rhizobia. In these experiments the number of bacteria added to each

treatment was identical; the variable was the increasing quantity of

soil. The six treatments used are listed below: Treatment 1 served as a non-

soil control; each treatment was run in duplicate.

The following soils were tested: Burabod, Clarion, Hubbard, Lualualei,

Molokai, Wahiawa, Waimea. Both the TAL-620 and Hawaii-5-0 strains were used.

The cultures were adjusted to approximately 2 x 107 cells/ ml with a Petroff-

Hauser counting chamber; one ml was added to the non-sterile soils and

incubated for two hours at room temperature (240C). The number of cells

contained in the inoculum was verified by IF-membrane-filter counts. The

influence of the various treatments, i.e. increasing soil, on the recovery of

strains was assayed by SRP (see Appendix Table 21).

Use of Different Diluents/Extractants and Flocculants to Increase Recovery of TAL-620 from Wahiawa Soil

The Wahiawa soil was chosen as the model problem soil. It consistently

gave one of the poorest recoveries. In these experiments different extracting

solutions were substituted for the H2O-Tween 80 extractant normally used in

the SRP procedure. Mid-exponential phase cultures of TAL-620 were enumerated

by Petroff-Hauser counts and adjusted to give 3 x 106 to 1 x 107 cells/gram of

non-sterile soil. The final inoculum size was confirmed by FA-filter counts.

The inoculated soils incubated two hours at ambient temperature (240C) prior

to assay for recovery by SRP.

The nonionic detergent Nonidet P40, (0.5% solution) was substituted for

Tween 80. Solutions of salts at different concentrations were tried (1M KCl,

3M NaCl), low pH (0.4M HCl), organic solvents (100% methanol, 10% ethanol),

two different anion extractants CuSO4: Ag2SO4 (used for NO3- and NO2- extraction

from soils, Jackson, 1962), and NaHCO3 (used for the extraction of soil

phosphorus, Jackson, 1962), solutions of the chelator EDTA (0.01 and 0.1M),

two solutions of peptone (1%, 2%), and two solutions containing partially

hydrolyzed gelatin (PHG) were also substituted.

To prepare PHG a 1% solution of Bacto-Gelatin was adjusted to pH 10.3

with 1N NaOH and autoclaved for ten minutes at 1210C to partially hydrolyze

the gelatin. When required the 1% solution was diluted to 0.1% either with

water, or with 0.1M phosphate buffered saline (PBS) (see Appendix Table 22);

thimerosal was added (final concentration 1:10,000) to prevent bacterial

growth in all gelatin solutions.

Results and Discussion

Reports in the literature indicate that the soil release procedure (SRP)

(see Appendix Table 21) does not fully release and extract Rhizobium and other

bacteria from temperate soils (Reyes and Schmidt, 1979; Vidor and Miller,

1979a; Schmidt, 1974). In addition, seven soils representative of those

commonly found in the tropics were highly sorptive for strains of fast-growing

rhizobia when assayed by this procedure. The recovery of TAL-620, a strain of

chickpea Rhizobium, from eight tropical soils is summarized in Table 6. All

soils except one (Waimea) were highly sorptive for the added bacteria;

recoveries were extremely low, in several cases less than 1% of the added

level.

The Wahiawa Oxisol was the most highly sorptive of the eight soils and

gave the poorest recovery. Poor recovery of Rhizobium was not related to soil

order as both Oxisols (highly weathered, “old” soils) and Inceptisols (little

weathered, “young” soils) retained the bacteria. (The soil order is the

broadest level in soil classification and is equivalent to the phylum in

biological classification.) Recovery of bacteria from soil may be related to

the finer classification of soils such as at the sub and great group level,

where both mineralogy and soil forming conditions are considered: from a typic

eutrandept (Waimea) 100% of the inoculated bacteria were recovered, while

recovery from hydric dystrandepts (PLP, LPHS, Burabod) and an andic ustic

humitropept (Makiki) were much poorer.

The capacity of 5 tropical and 2 temperate soils to sorb bacteria and

their recovery and enumeration by SRP was shown by the soil titration

experiments. Figures 6, 7, and 8 show the inverse relationship between

increasing soil content and recovery; as the quantity of soil in each

treatment increased, recovery of added Rhizobium decreased. The strain of

lentil Rhizobium Hawaii-5-0 was strongly sorbed by the Wahiawa soil as shown

in Figure 6. This indicates that the poor recovery of TAL-620 was not a

strain specific phenomenon. In fact Hawaii 5-0 was more strongly adsorbed by

the Wahiawa soil than TAL-620. The Waimea Inceptisol (see Figure 7) and the

two midwestern soils (Figure 8) had less sorptive capacity relative to the

other soils; the decrease in recovery of added bacteria was not as great.

Seventeen different extractants were substituted for H2O in the SRP

recovery procedure, these included 2 nonionic detergent, salts, low pH,

alcohols, solutions of peptone, EDTA, and two solutions of partially

hydrolyzed gelatin. Of these 17 extractants (see Tables 7 and 8) four gave

recoveries greater than 30% (see Table 8): 0.5M NaHCO3 (33% recovery), 1%

peptone (33% recovery), and 0.1% PHG in phosphate buffered saline (60%

recovery). It was important to select the proper concentration for the

Figure 6. Soil Titrations - recovery of TAL-620 (chickpea

Rhizobium) from two Hawaiian Oxisols, Molokai (-●--●-) and Wahiawa (-○--○-), and recovery of Hawaii-5-0 (R. leguminosarum) from Wahiawa (-◘--◘-) using Soil Release Procedure (SRP) methodology— both strains strongly affected by increasing percent of soil--the greater the content of soil the fewer cells recovered.

Figure 7. Soil Titrations - recovery of TAL-620 from 3 tropical

soils Waimea (-◘--◘-), Lualualei (-●--●-), and Burabod (-○--○-) using Soil Release Procedure (SRP) methodology. The Waimea Inceptisol has less sorptive capacity than the other soils.

Figure 8. Soil Titrations - recovery of TAL-620 from two mid-

western Mollisols Clarion (-●--●-), and Hubbard (-○--○-) using soil release procedure methodology. These soils are less strongly sorptive than most of the tropical soils tested.

extractant solution. In three cases (see Table 8) use of higher

concentrations of the extractant gave poorer recovery than with the less

concentrated solution. This was also true of PHG solutions. In preliminary

experiments with solutions of partially hydrolyzed gelatin greater than

0.1%, the soil colloids would not flocculate out of suspension. Less than

0.1 ml of the solution could be filtered. Because of this, PHG was used at

a final strength of 0.1%. The composition of the solution with which the 1%

partially hydrolyzed gelatin was diluted, to give the final 0.1% working

mixture, was important: gelatin in O.1M PBS gave higher recovery than

gelatin diluted in H20.

Niepold et al. (1979) proposed that the chemical composition of the

extraction fluid should influence recovery. My results, obtained with the

different extractants, support this proposal. However, nonionic detergents,

high ionic strength, low pH, and polar organic solvents were ineffective.

Organic extractants containing protein digests (peptone) or proteins

(gelatin) were the most successful. The extractants containing PHG gave the

best results. Therefore the failure to recover rhizobia from tropical soils

was a function of inadequate extraction of the bacteria; recovery was

increased simply by altering the chemistry of the extraction fluid.

Those extractants with complex charge chemistries, such as solutions

of PHG, led to the extraction of more bacteria. In fact, PHG solutions in

phosphate buffered saline gave the greatest recovery. This might indicate

that ion exchange processes (Na+) as well as charge neutralization (PO43-

ions and charges of PHG molecules) are important in the recovery of bacteria

from the soil matrix. The phosphate ions might aid in neutralizing the

positive charges at the broken edges of the clay plates; the sodium ions,

due to mass action can displace divalent cations, bridging bacteria to soil

colloids, or bridging clay particles together and prevent adequate

dispersion. The gelatin ions may also combine and neutralize charges within

the soil matrix. Since soils are highly complex ion exchange systems it

seems logical that a solution with a complex charge status is required to

recover the bacteria. If cation exchange is an important phenomenon in the

recovery of bacteria from soils, a cation with more exchange power such as

K+ or NH4+ might lead to increased recovery of Rhizobium.

CHAPTER 5

MODIFIED MEMBRANE FILTER - IMMUNOFLUORESCENCE FOR ENUMERATION OF RHIZOBIUM FROM TROPICAL SOILS

Introduction

Immunofluorescence (IF) provides a direct method for in situ autecological

studies of microorganisms; it allows for the simultaneous detection and

identification of the desired organism in its natural habitat. The technique

can be made quantitative by separating the bacteria from soil particles and

concentrating them on non-fluorescent membrane-filters for IF enumeration (Soil

Release Procedure, SRP, see Appendix Table 21). In applying SRP to study

Rhizobium in tropical soils I encountered great difficulty in releasing bacteria

from soil particles and recovering them for IF enumeration (see Chapter 2).

Most tropical soils were highly sorptive for Rhizobium when assayed by

SRP. However, as discussed by Niepold et al. (1979) the chemical composition of

the extraction fluid greatly influenced the degree of sorption. Extraction

solutions containing proteins (Partially Hydrolyzed Gelatin, PHG) were most

successful in increasing recovery of rhizobia from a Hawaiian Oxisol, chosen as

the model problem soil. Therefore, the failure to recover rhizobia sorbed to

soil is directly related to the techniques used to extract and enumerate them.

The bacteria are not irreversibly bound. All that is required is the proper

methodology to recover the cells. In addition, results obtained with partially

hydrolyzed gelatin solutions indicated that ion exchange may be important in the

recovery of bacteria from tropical soils.

The present work describes efforts to optimize recovery of Rhizobium from

tropical soils. Experiments to select the best diluting solution to use with

PHG were undertaken. In the course of these studies I realized that blending

the soils (SRP, Appendix Table 21) especially the well aggregated Oxisols led to

a breakdown of the stable aggregates and to a very large increase in surface

area. This increased surface, by creating more area for sorptive interactions,

probably led to lower recoveries. I reasoned that the chemical extractants

might work better when used with a less disruptive dispersion method. The final

outcome of these investigations was the development of a modified soil release

procedure (MSRP). The soils were dispersed by shaking on a wrist-shaker in

flasks containing glass beads and a chemical extractant using PHG.

Materials and Methods

Source and Maintenance of Cultures

Two strains of fast growing rhizobia, TAL-620 (Rhizobium for chickpea, see

Chapter 1, Table 1), Hawaii-5-0 (Rhizobium leguminosarum, May, 1979), and two

strains of slow-growing rhizobia, USDA 31 and USDA 110 (both R. japonicum) (Dr.

D. F. Weber, USDA Beltsville, Md.) were used in these experiments. All strains

were maintained on a modified YEMS medium (see Appendix Table 19a), and when

required were grown in broth of the same composition. Exponential phase

cultures were enumerated by direct counts with a Petroff-Hauser chamber. The

cultures were diluted in saline to the desired inoculum size. For experiments,

one or two ml of the inoculum was added to moistened soil. Strict aseptic

techniques were followed at all steps when autoclaved soils were to be

inoculated.

Chemical Reagents

The chemical reagents have already been described (see Chapter 2,

Materials and Methods). Granulated gelatin was obtained from three sources:

Difco Bacto-Gelatin (2 lots, control #459821, and #464194), Difco, Detroit,

Michigan; Fisher Bacterilogical Gelatin (Lot number illegible), Fisher

Scientific, Fairlawn, N. J.; U.S.P. Granular Gelatin (No lot number), Pioneer

Chemical Co., Inc., Long Island City, N. Y.

Soil Samples, Soil Sterilization

The soils used in these experiments, and their method or preparation and

distribution were described previously (see Chapter 2, Materials and Methods).

Soil sterilization, when needed was done by autoclaving for 1.5 hours at 1210C

(Bohlool, 1971).

Preparation of Fluorescent Antibodies, FA Staining, Epifluorescence Enumeration

Fluorescent antibodies were prepared by the method of Schmidt

et al. (1968). The soil release enumeration procedure (SRP) (Bohlool and

Schmidt, 1973a; Schmidt, 1974) is described in Appendix Table 21). Modifications

to the published procedures were described previously (see chapter 2, Materials

and Methods).

Preparation of Partially Hydrolyzed Gelatin (PHG)

A 1% solution of gelatin (10 g/L H20) was adjusted to pH 10.3 with 1N

NaOH, and autoclaved 10 minutes at 1210C to partially hydrolyze the gelatin.

The hydrolyzed solution was stored at 40C until used; thimerosal was added,

1:10,000 final concentration, as a preservative. All experiments unless

indicated otherwise used Difco Bacto-Gelatin Lot No. 464194 (will be referred to

as gel #1).

SRP - Effect of Different Strength Gelatin Solutions

To determine what concentration of gelatin would give maximum recovery of

added bacteria, a 1% solution of PHG was used either undiluted (1%) or was

diluted with water to give solutions containing 0.01%, 0.05%, 0.08%, 0.1%, 0.15%

PHG. All solutions were adjusted to pH 7, and thimerosal was added, 1:10,000,

to prevent bacterial growth. A mid exponential phase culture of TAL-620 was

enumerated by direct counts with a Petroff-Hauser chamber, and adjusted to

1 x 107 cells/ml; one ml portions were inoculated into 25 x 200 mm screw-cap

tubes containing 10 g of non-sterile Wahiawa soil. Duplicate tubes were

inoculated for each gelatin concentration tested (12 tubes). The soils, after

incubating for two hours at 240 C, were assayed with SRP for recovery of

inoculated Rhizobium. The procedures were the same as described in Appendix

Table 21 except that the various PHG solutions were substituted for H2O-Tween

80.

SRP - Recovery of TAL-620 from Wahiawa Soil, 0.1 PHG (in H2O) at Different pH’s

Solutions of 0.1% PHG (1% diluted to 0.1% with H2O) were adjusted to pH 4,

6, 7, 8, and pH 9 with 1N HCl and 1N NaOH. Thimerosal was added, 1:10,000, to

prevent bacterial growth. A mid exponential phase culture of TAL-620 was

enumerated by direct counts with a Petroff-Hauser chamber, and adjusted to 1 x

107 cells/ ml. One ml was inoculated into 10 g of non-sterile Wahiawa soil to

give approximately 1 x 106 cells/g soil; duplicate tubes were used for each PHG

solution. The inoculated soils were incubated for two hours at 24oC prior to

assay for recovery. The assay procedure was the same as that in Appendix Table

21 except that PHG solutions were substituted for the H2O-Tween 80 extractant.

SRP - Recovery of TAL-620 from Wahiawa Soil, 0.1% PHG, Use of Different Diluents

A 1% solution of PHG was prepared as described previously. After

autoclaving, the solution was adjusted to pH 7 and thimerosal was added

1:10,000. The following solutions were used to test for percentage of recovery

in conjunction with PHG (final strength of PHG - 0.1%): H20; 0.1M Na-EDTA; 0.5M

NaHCO3; 0.001M NaHMP; 0.1M PBS. A mid exponential phase culture of TAL-620

enumerated, as described above, was adjusted to 1 x 107 cells/ml; one ml was

inoculated into 10 g of non-sterile Wahiawa soil to yield approximately 1 x 106

cells/g of soil; duplicate tubes were used per each PHG solution tested. After

a two hour incubation at 24oC the soils were assayed for recovery with SRP.

Procedures were the same as those described in Appendix Table 21 except that PHG

solutions were substituted for H2O-Tween 80.

Development of a Modified Soil Release Procedure (MSRP)

Rather than blending the soil, the modified procedure required shaking the

soil and extracting solution together on a wrist-action shaker, in a screw-cap

flask with glass beads. Several PHG-diluent combinations were tested: 0.1% PHG

- H20; 0.1% PHG - O.1M Na2HPO4 (pH 9); 0.1% PHG - 0.1M (NH4)2HPO4 (pH 8.3); 0.1%

PHG - 0.1M PBS (pH 7.1). Thimerosal was added to all gelatin solutions, final

strength 1:10,000. A mid exponential phase culture of TAL-620 was adjusted to 1

x 107 cells/ml as described previously; one ml was inoculated into 10 g of non-

sterile Wahiawa soil to give approximately 1 x 106 cells/g of soil. Duplicate

tubes were used for each PHG solution tested. The inoculated soils were

incubated for two hours at 24oC prior to assay for recovery.

The assay procedure was the same as that described in Appendix Table 21

with the following modifications: the soils and extracting solution were added

to 250 ml screw cap flasks containing 25 - 30 g of 3 mm glass beads. The flasks

were shaken for five minutes on a Burrel wrist-action shaker to disperse the

soil, all other procedures (flocculation, filtration, enumeration, etc.) were

the same as those described in Appendix Table 21.

MSRP - Effect of the Decreasing Hydrated Radius of Four Monovalent Cations Upon Recovery of TAL-620 from Wahiawa Soil

Solutions of 0.2M strength prepared from LiCl, NaCl, KCl and NH4Cl were

mixed with 1% PHG (prepared as described previously) to yield the following

extractant solutions: 0.1% PHG - 0.2M LiCl; 0.1% PHG - 0.2M NaCl; 0.1% PHG -

0.2M KCl; 0.1% PHG - 0.2M NH4Cl. A mid exponential phase culture of TAL-620 was

adjusted to 3 x 107 cells/ml as described previously; one ml was added to 10 g

of non-sterile Wahiawa soil, in duplicate tubes. After a two hour incubation at

24oC the soils assayed for recovery by MSRP.

MSRP - Effect of Gelatin Type (Manufacturer) and Shaking Time on Recovery of TAL-620 from Wahiawa Soil

Four 1% solutions were prepared from each of the granulated gelatin

preparations listed above under Chemical Reagents; thimerosal was added as a

preservative and the solutions were stored at 40C until used. All gelatin

solutions were adjusted to 0.1% PHG with O.1M (NH4)2HPO4 (pH 8.3). A mid

exponential phase culture of TAL-620 was adjusted to 2 x 106 cells/ml, one ml

was inoculated into duplicate tubes containing 10 g of non-sterile Wahiawa soil.

A two hour incubation at 240C preceeded the recovery assay. The soils and

extractants were shaken for five minutes.

The effect of shaking time was evaluated using Difco gel #1. A 1% PHG

solution was prepared as described previously; this solution was adjusted to

0.1% with O.1M (NH4)2HPO4 (final pH of extractant = 8.3). Four dispersion times

were tried. The soils and extractant were shaken for 5, 15, 30, and 60 minutes.

MSRP - Soil Titrations

The procedures used to construct the titrations were described in the

Materials and Methods section of Chapter 2. The SRP assay discussed in Chapter

2 and this assay (MSRP) were conducted simultaneously. For these experiments

the MSRP extractant consisted of 0.1% PHG - O.1M (NH4)2HPO4 (final pH = 8.3);

the soil and extracting solution were shaken for five minutes.

Growth of Fast and Slow Growing Rhizobium in Sterile Wahiawa Soil

The growth of TAL-620 was followed for 20 days in sterile Wahiawa soil.

The bacteria were enumerated directly from the soil by both viable counts and

MSRP (MSRP = 0.1% PHG - O.1M (NH4)2HPO4, shaking for 5 minutes). One set of

data was analyzed by viable counts, SRP and MSRP. For viable counts soils were

mixed with 95 ml of sterile H2O in sterile screw cap flasks containing 25 - 30 g

of 3 mm glass beads, and shaken for five minutes. After dispersion, the

appropriate dilutions were plated onto YEMS medium (see Appendix Table 19a)

using the Miles and Misra drop plate technique (Vincent, 1970). The procedure

was modified to use a Pipetman P-20 adjustable micropipeter (Rainin Instrument

Co., Inc., Woburn, Mass.) and disposable tips (sterilized by autoclaving). The

pipetor was adjusted to dispense 20 pl of solution; one disposable tip was used

per dilution.

A mid exponential phase culture of TAL-620 was adjusted to 1 x 106

cells/ml; and one ml was inoculated into 10 g of sterile Wahiawa soil to yield

approximately 1 x 105 cells/g. When necessary the soils were moistened with

several drops of sterile distilled water. The tubes were sampled randomly, four

tubes were assayed per data point.

The growth of two strains of R. japonicum was followed for seven days

(USDA 31) and fourteen days (USDA 110) in sterile Wahiawa soil. The bacteria

were enumerated by viable count (as described), SRP, and MSRP. Mid exponential

phase cultures of USDA 31 and USDA 110 were enumerated in a Petroff-Hauser

chamber, adjusted to approximately 105 cells/ml, and one ml was aseptically

inoculated into the soils to give approximately 104 cells/g. Population levels

were determined after five days incubation at 280C (USDA 31, USDA 110), seven

days (USDA 31), and fourteen days (USDA 110). Duplicate tubes were used for

each measurement.

Growth of USDA 110 in Sterile Clarion Soil

The growth of USDA 110 was followed in sterile Clarion soil for 4 days

using plate counts, SRP and MSRP. A mid exponential phase culture of USDA 110

was enumerated with a Petroff-Hauser chamber and adjusted to 2 x 106 cells/ml;

one ml was inoculated aseptically into tubes containing 10 g of autoclaved

Clarion soil to give approximately 2 x 105 cells/g. The cells were counted

three and five days after inoculation. The tubes were incubated at 280C;

duplicate tubes were used for each measurement.

Statistical Methods

A one way analysis of variance (ANOVA) (Sokal and Rohlf, 1969) was used to

determine if the monovalent cation treatments had a significant effect on

extraction of Rhizobium from soils and to determine if the type of gelatin was

important, as well as the shaking/extraction period.

Results and Discussion

The results discussed in Chapter 2 indicated the difficulty in

quantitatively recovering bacteria from tropical soils when using SRP (see

Appendix Table 21); seven of eight tropical soils were highly sorptive for added

rhizobia. Several other investigators have also experienced difficulty in

recovering bacteria quantitatively from soil and sediments (Vidor and Miller,

1979a; Reyes and Schmidt, 1979; Reed and Dugan, 1978). However, the

substitution of solutions of partially hydrolyzed gelatin (PHG) led to increased

recovery of added bacteria.

Solutions of gelatin, due to their protein origins are amphoteric, i.e.

they have pH dependent charges. The charges on the gelatin molecules can

interact with both soil colloids and microbial cells. Partial hydrolysis of the

gelatin increases the number of small peptides which can interact with the soils

exchange complex and therefore satisfy charges within the matrix. By satisfying

charges on both bacteria and soil particles the two should not bind due to any

charge interactions. Preliminary evidence indicated that ion exchange phenomena

might be involved in the extraction of bacteria from the soil matrix. PHG

solutions mixed with phosphate (sodium) buffered saline gave greater recovery

than PHG mixed with H2O (see Table 8). Therefore, experiments were designed to

establish the optimum concentration of PHG; to establish the importance of the

extracting solution, and to determine the proper combination of ions for

increasing recovery, as well as to determine if ion exchange processes were

involved.

A solution of 0.1% PHG gave the best quantitative recovery of TAL-620 from

the Wahiawa soil (see Table 9). Concentrations of gelatin greater than 0.1%

prevented flocculation of the soil colloids. I did not determine if the lack of

flocculation was due to the increased viscosity of the more concentrated PHG

solutions, relative to H2O or to an increase dispersive effect. The addition of

more flocculant did not precipitate the soil colloids.

In addition to selecting the proper PHG concentration, it was necessary to

select the optimum pH for extraction (see Table 10). A low pH gelatin solution

(pH 4) gave poorer quantitative recovery than solutions at higher pH. Solutions

at pH 8 and pH 9 gave quantitative recovery of added Rhizobium. However, good

recovery was obtained at pH 6 and pH 7. Preliminary results had demonstrated

that high pH would disperse the Wahiawa Oxisol. However, high pH alone did not

lead to quantitative recovery of Rhizobium (Table 8, NaHCO3 solutions, pH 8.2 -

8.3).

Previous results (see Chapter 2, Table 8) demonstrated the importance of

selecting the proper diluent to mix with the gelatin solution, e.g. H2O vs. PBS.

I decided to mix several of the solutions which had given slightly increased

recoveries when substituted for H2O-Tween 80 in SRP with solutions of PHG. Both

sodium-EDTA and sodium bicarbonate solutions were not satisfactory when mixed

with PHG (see Table 11) as recoveries were below 50% of the inoculum. A

combination of PHG and the soil deflocculant sodium hexa-meta-phosphate did not

increase recovery of added Rhizobium. In addition, extractant solutions of PHG

and water gave quite variable results from 20% recovery up to 80%.

During the course of these experiments, trends in the data indicated that

there was a difference the blendor cups. The percentage recovery obtained

from extractions in one blendor cup were consistently higher than those

obtained from another cup. However, the level of recovery could be

decreased simply by installing new bearing and blade assemblies in the cups.

Apparently, the bearings in the older blade assembly slowed the blending

speed, and greater recoveries were obtained than with blendors with new

assemblies. The decrease in recovery might have been a function of

increased soil area brought about by high speed blending. Oxisols are well

aggregated soils; however, the aggregates are composed of small clay size

particles that bind together (Sanchez, 1976). Strong blending results in

high shear forces that destroy the aggregates, consequently leading to a

very large increase in surface area. This increased surface area can

interact with microorganisms and bind or trap them in floccules.

I decided that a modified soil dispersion method was needed; one that

did not fully deaggregate the soils yet also gave consistently good

recovery. The procedure finally adopted was shaking the soils and

extracting solutions with a wrist-action shaker, employing glass beads to

increase agitation. Four 0.1% PHG-diluent combinations were tested (Table

12). The combination offering the most promise was 0.1% PHG - 0.1M

(NH4)2HPO4 (pH 8.3).

These experiments gave more concrete evidence that an ion exchange-

type of process might be important; poorer recoveries were obtained with

sodium salt - PHG combinations than with ammonium salt - PHG combinations.

Sodium is known to be a poorer exchanger than ammonium (Brady, 1974). For

this reason I decided to see if the cations themselves could influence

recovery. The results in Table 13 show that the cation does have a

significant influence on recovery (F = 33, F.01(3,22) = 4.87). It may also

be inferred that the phosphate radical is important, as greater recoveries

were obtained with ammonium and sodium phosphate-PGH combinations than with

their corresponding chloride salts.

The PHG extracting solution adopted as the standard MSRP extractant

consisted of 0.1% PHG - 0.1M (NH4)2HP04, pH 8.3. The extraction time was not

important (see Table 14) and a five minute extraction was adopted as the

standard. However, the type of gelatin (manufacturer) used in the PHG

solutions was important (see Table 15). The Difco Bacto-Gelatins (Gels 1 and

2) gave the best recovery. Granulated gelatins are prepared by both acidic and

basic digests of animal biproducts. This results in different products. In

addition the starting materials composing any one lot of granulated gelatin

might vary and this can influence the final product. Even though the pH of the

four solutions varied prior to hydrolysis (pH Difco gel 1 and 2 6.2, pH Fisher

4.6, pH Pioneer 7.6) each solution had an identical pH after hydrolysis and

mixing with O.1M (NH4)2HPO4 (pH 8.3). The differences in recovery were

therefore not due to any pH effects. Due to the differences in the granulated

gelatins several different lots and manufacturers should be tested. The lot

giving the greatest recovery should then be used.

The MSRP procedure finally adopted as the standard method for the

extraction of Rhizobium from tropical soils is described in Table 16. The

major difference between this procedure and the SRP method is the extraction

step, a chemical extractant is employed, and the soil is dispersed by rapid

shaking rather than blending. All other procedures are the same as those

described in Appendix Table 21.

The efficiency of the gelatin technique compared to SRP is illustrated in

Figures 9, 10 and 11. The SRP technique was greatly affected by soil content

and did not possess the ability to desorb the bacteria; therefore, as the soil

content increased the number of bacteria recovered decreased. The gelatin MSRP

method was hardly affected by the amount of soil and was able to desorb and

extract the added Rhizobium. The Waimea Inceptisol (see Figure 10) and the

other two Midwestern soils (Figure 11) were less sorptive for rhizobia than the

Figure 9. Soil Titrations - comparison of the Modified Soil

Relese Procedure (MSRP) (triangles)and the Soil Release Procedure (SRP) (circles) for the recovery of Rhizobium from tropical soils. The MSRP Procedure (two Oxisols) is not affected by soil content, while recovery with the SRP procedure is soil dependent. (SRP data same as in Figure 6).

Figure 10. Soil Titrations - comparison of the Modified Soil

Release Procedure (MSRP) (triangles) and the Soil Release Procedure (SRP) (circles) for the recovery of a strain of chickpea Rhizobium from tropical soils. The MSRP procedure is not affected by soil content, while recovery with the SRP procedure is soil dependent. (SRP data same as in Figure 7)

Figure 11. Soil Titrations - comparison of the Modified

Soil Release Procedure (MSRP) (triangles) and the Soil Release Procedure (SRP) (circles) for the recovery of a strain of chickpea Rhizobium from two midwestern Mollisols. The MSRP procedure is not affected by soil content, while recovery with the SRP procedure is more soil dependent. (SRP data same as in Figure 8)

other tropical soils when the SRP assay was employed. This was discussed

previously in Chapter 2. Even though these three soils were less sorptive,

the MSRP procedure was able to recover more Rhizobium cells.

Bacteria growing in soil may attach to soil particles by different

mechanisms than bacteria that are added from culture and extracted shortly

thereafter. Scheraga et al. (1979) found that bacteria added to marine

sediments were rapidly sorbed. The sorption phenomenon followed a Langmuir

plot (Fletcher, 1977) and therefore indicated physico-chemical adsorption.

Although Marshall (1980) stated that no definite evidence exists that would

indicate polymeric bridging as important in the binding of bacteria to soils

and sediments, polymer exudations from bacteria could bind them to soil

particles.

In the previous experiments, where attempts to modify the soil release

procedure (SRP) (see Appendix Table 21) resulted in a modified method of

dispersion, the bacteria were added to the soil and recovered two hours

later. Since the bacteria were not growing in the soil, physico-chemical

binding was predominant. The MSRP method was successful in overcoming these

forces and extracting the bacteria from the various soils. To determine the

feasibility of using MSRP for studies of Rhizobium growing in nature, where

polymeric briding might be important, several sterile systems were

established. The advantage of the sterile systems is that it allows the

various IF enumeration procedures to be checked against viable counts. Under

non-sterile conditions, as discussed previously, it is extremely difficult to

enumerate Rhizobium by plate counts.

The growth of TAL-620 was followed for 20 days in sterile Wahiawa

Oxisol by plate counts and the modified soil release procedure. One set of

data was analyzed by plate counts, the soil release procedure and the

modified soil release procedure (see Figure 12).

The MSRP method underestimated the plate counts at the lower numbers of

Figure 12. Growth of TAL-620 in sterile Wahiawa Oxisol followed

by plate counts (-●--●-), the modified soil release procedure (-▲--▲-), and the soil release procedure ( ■ ) (one data point only).

cells and overestimated at the higher numbers. Bohlool and Schmidt (1973a)

observed a similar phenomenon when the growth of USDA 110, a strain of R.

japonicum, was followed in a sterile midwestern soil. They attributed the

overestimation of cells by the IF enumeration procedure (SRP) to the

accumulation of dead bacteria. In a natural environment dead cells would be

degraded fairly rapidly by other organisms present in the environment.

In contrast to the relatively close correlation between the MSRP-IF

enumeration procedure and plate counts, the SRP procedure and viable counts

did not correlate. While both plate counts and viable counts indicated a

population level of approximately 107 cells/g of soil the SRP method

indicated only 103/g. The MSRP technique, as well as being able to

quantitatively extract bacteria added to soil was able to quantitatively

extract Rhizobium growing in soil.

Marshall (1969a,b) has shown that slow growing rhizobia have a

different surface charge character than the fast growing rhizobia. The

differences in surface charge might affect the ease with which slow and fast

growing rhizobia are extracted from tropical soils. To ensure that the MSRP

method was satisfactory for the enumeration of both fast and slow growing

rhizobia two strains of R. japonicum were inoculated into sterile Wahiawa

soil. The growth of USDA 31 and USDA 110 was followed by plate counts, the

soil release procedure (SRP), and the modified soil release procedure (MSRP)

(see Table 17). Although the MSRP method generally recovered more cells

than the SRP method, neither procedure approached the levels indicated by

the viable count. This problem is most likely soil dependent, however.

When USDA 110 was followed in sterile (Table 18) Clarion soil (the

soil:strain combination used to develop SRP) there was very close agreement

between the viable count and the population level indicated by the modified

procedure. However, in contrast to what I expected, the SRP method did not

correlate closely with the plate counts or with MSRP. This is rather

puzzling as close correlation was reported in the literature (Bohlool and

Schmidt, 1973a). However, the problem may be related to the use of blendors,

as described earlier.

In general the modified soil release procedure was able to recover

Rhizobium from soils more efficiently than the previously published

immunofluorescence enumeration procedure (SRP, Appendix Table 2l). Most likely

the success of the modified method in recovering bacteria was due to the use of

a complex, multicomponent chemical extractant (0.1% PHG - O.1M (NH4)2HPO4) and a

method of dispersion that did not create high shear forces such as occur in

blending. The lack of high shear force is important when working with Oxisols

since the well aggregated subunits will disintegrate into clay sized components

if subjected to too much stress. This can lead to a large increase in the

surface area available for microbes to interact with the clays. Without a

complex chemical extractant to reduce the number of charges within the soil

matrix, it seems likely that the bacterial cell would become rapidly sorbed to

clay particles, thus reducing the level of recovery.

Ion exchange may be an important phenomenon in the extraction and

recovery of bacteria by the MSRP procedure. However, more experiments need to

be performed under carefully controlled conditions to fully elucidate the

relationship of cation (and perhaps anion) exchange to recovery of Rhizobium

when using MSRP. For the present, however, the method successfully extracts

Rhizobium from tropical soils more efficiently than previously published

procedures. This should not permit the study of the ecology of free-living

Rhizobium in tropical soils.

APPENDICES

LITERATURE CITED Advisory Committee on Technology Innovation. 1979. Tropical

Legumes: Resources for the Future. National Academy of Sciences. Washington, D.C.

Alexander, M. 1965. Most-probable number method for microbial

populations. In C. A. Black (Ed.) Methods in Soil Analysis. American Society of Agronomy. Madison. pp. 1467-1472.

Balkwill, D. L., D. P. Labeda and E. L. Casida, Jr. 1975. Simplified

procedures for releasing and concentrating microorganisms from soil for transmission electron microscopy viewing as thin- sectioned and frozen-etched preparations. Can. J. Microbiol. 21:252-262.

Barber, L. E. 1979. Use of selective agents for recovery of

Rhizobium meliloti from soil. Soil Sci. Soc. Am. J. 43:1145-1148. Belly, R. T., B. B. Bohlool and T. D. Brock. 1973. The genus

Thermoplasma. Ann. N. Y. Acad. Sci. 225:94-107. Berger, J. A., S. N. May, L. R. Berger and B. B. Bohlool. 1979.

Colorimetric enzyme-linked immunosorbant assay for the identification of strains of Rhizobium in culture and in nodules of lentils. Appl. Environ. Microbiol. 37:642-646.

Bohlool, B. B. 1971. Immunofluorescence approach to the study of

Rhizobium japonicum in soils. Ph.D. Dissertation. University of Minnesota. Minneapolis.

__________ and E. L. Schmidt. 1968. Nonspecific staining:

its control in immunofluorescence examination of soil. Science. 162:1012-1014.

__________ and __________. 1970. Immunofluorescent detection

of Rhizobium japonicum in soils. Soil Sci. 110:229-236. __________ and __________. 1973a. A fluorescent antibody

technique for determination of growth rates of bacteria in soil. Bull. Ecol. Res. Comm. (Stockholm). 17:336-338.

__________ and __________. 1973b. Persistence and competition

aspects of Rhizobium japonicum observed in soil by immunofluorescence microscopy. Soil Sci. Soc. Am. Proc. 37:561-564.

__________ and __________. 1979. The immunofluorescence

approach in microbial ecology. Advances in Microbial Ecology Vol. IV. In Press.

Brady, N. C. 1974. The Nature and Properties of Soils. Macmillan

Publishing, Co., Inc. New York. Brock, T. D. 1971. Microbial growth rates in nature. Bacteriol.

Rev. 35:39-58.

Brockwell, J., E. A. Schwinghamer and R. R. Gault. 1977. Ecological studies of root-nodule bacteria introduced into field environments. V. A critical examination of the stability of antigenic and streptomycin-resistance markers for the identification of strains of Rhizobium trifolii. Soil Biol. Biochem. 9:19-24.

Broughton, W. J. and M. J. Dilworth. 1971. Control of leghaemoglobin

synthesis in snake beans. Biochem. J. 125:1075-1080. Bushnell, 0. A. and W. B. Sarles. 1939. Investigations upon the

antigenic relationships among the root-nodule bacteria of the soybean, cowpea and lupine cross-inoculation groups. J. Bacteriol. 38:401-410.

Butler, J. N. 1964. Ionic Equilbria, A Mathematical Approach.

Addison-Wesley Publishing Co. Palo Alto. Caldwell, B. E. and G. Vest. 1968. Nodulation interactions between

soybean genotypes and serogroups of Rhizobium japonicum. Crop Sci. 8:680-688.

Coons, A. H., H. J. Creech, R. N. Jones and E. Berliner. 1942.

Demonstration of pneumococcal antigen in tissue by the use of fluorescent antibody. J. Immunol. 45:159-170.

Corbin, E. J., J. Brockwell and R. R. Gault. 1977. Nodulation

studies on chickpea (Cicer arietinum). Aust. J. Exp. Agr. Anim. Husb. 17:126-134.

Damirgi, S. M., L. R. Frederick and I. C. Anderson. 1967. Serogroups

of Rhizobium japonicum in soybean nodules as affected by soil types. Agron. J. 59:10-12.

Daniels, S. L. 1972. The adsorption of microorganisms onto solid

surfaces: A review. Dev. Ind. Microbiol. 13:211-253. Date, R. A. and A. M. Decker. 1965. Minimal antigenic constitution

of 28 strains of Rhizobium japonicum. Can. J. Microbiol. 11:1-8. __________ and J. M. Vincent. 1962. Determination of the number of

root-nodule bacteria in the presence of other organisms. Aust. J. Exp. Agric. Anim. Husb. 2:5-7.

Dazzo, F. B. and D. H. Hubbel. 1975. Antigenic differences between

infective and noninfective strains of Rhizobium trifolii. Appl. Microbiol. 30:172-177.

Diatloff, A. 1977. Ecological studies of root-nodule bacteria

introduced into field environments. 6. Antigenic and symbiotic stability in Lotononis rhizobia over a 12 year period. Soil Biol. Biochem. 9:85-88.

Dudman, W. F. 1964. Immunodiffusion analysis of the extracellular

soluble antigens of two strains of Rhizobium meliloti. J. Bacteriol. 88:782-794.

__________. 1971. Antigenic analysis of Rhizobium japonicum by immune diffusion. Appl. Microbiol. 973-985.

__________. 1977. Serological methods and their application to

dinitrogen fixing organisms. In R. W. F. Hardy and A. H. Gibson (Eds.). A Treatise on Dinitrogen Fixation. Section IV: Agronomy & Ecology. John Wiley & Sons. New York. pp. 487-508.

__________ and J. Brockwell. 1968. Ecological studies of root-

nodule bacteria introduced into field environments. l. A survey of field performance of clover inoculants by gel immune diffusion. Aust. J. Agric. Res. 19:739-747.

Edwards, P. R. and W. H. Ewing. 1955. Identification of Enterobacteriaceae.

Burgess Publishing Co. Minneapolis. Eisen, H. N. 1974. Immunology: An introduction to molecular and

cellular principles of the immune responses. Harper & Row. New York.

Faegri, A., V. L. Torsvik and J. Goksoyr. 1977. Bacterial and

fungal activities in soil: separation of bacteria and fungi by a rapid fractionated centrifugation technique. Soil Biol. Biochem. 9:105-112.

Fletcher, M. 1977. The effects of culture concentration and age,

time and temperature on bacterial attachment to polystyrene. Can. J. Microbiol. 23:1-6.

__________. 1979. A microautoradiographic study of the activity

of attached and free-living bacteria. Arch. Microbiol. 122: 271-274.

Francis, C. W., W. P. Bonner and T. Tamura. 1972. An evaluation of

zonal centrifugation as a research tool in soil science. I. Methodology. Soil Sci. Soc. Amer. Proc. 36:366-376.

Fred, E. B., I. L. Baldwin and E. McCoy. 1932. Root-nodule Bacteria

and Leguminous Plants. University of Wisconsin Press. Madison. Gibbins, L. N. 1967. The preparation of antigens of Rhizobium

meliloti by ultrasonic disruption: an anomaly. Can. J. Microbiol. 13:1375-1379.

Graham, P. H. 1963. Antigenic affinities of the root-nodule bacteria

of legumes. Antonie van Leewenhock. 29:281-291. __________. 1969. Selective medium for growth of Rhizobium.

Appl. Microbiol. 17:769-770. Graham, P. H. and G. A. Parker. 1964. Diagnostic features in the

characterization of the root-nodule bacteria of legumes. Plant Soil. 20:383-396.

Hattori, T. 1973. (Ed.) In Microbial Life in Soil. An Introduction.

Marcel Dekker Inc. New York.

__________ and R. Hattori. 1976. The physical environment in soil microbiology: An attempt to extend principles of microbiology to soil microorganisms. Crit. Rev. Microbiol. 4:423-461.

Hobbie, J. E., R. J. Daley and S. Jasper. 1977. Use of nuclepore

filters for counting bacteria by fluorescence microscopy. Appl. Environ. Microbiol. 33:1225-1228.

Hornby, D. and A. J. Ullstrop. 1965. Physical problems of sampling

soil suspensions in the dilution-plate technique. Phytophathology. 55:1062. (Abstract)

Huges, D. Q. and J. M. Vincent. 1942. Serological studies of the

root-nodule bacteria. III. Tests of neighboring strains of the same species. Proc. Linnean Soc. N.S. Wales. 67:142-152.

Humphrey, B. A. and J. M. Vincent. 1965. The effect of calcium

nutrition on the production of diffusable antigens of Rhizobium trifolii. J. Gen. Microbiol. 41:109-118.

__________ and __________. 1969. The somatic antigens of

two strains of Rhizobium trifolii. J. Gen. Microbiol. 59:411-425. __________ and __________. 1975. Specific and shared

antigens in strains of Rhizobium meliloti. Microbios. 13:71-76. Ikawa, H. 1979. Laboratory data and descriptions of soils of the

Benchmark Soils Project. Vol. I. Hawaii Project. Hawaii Agricultural Experiment Station, University of Hawaii. Misc. Publ. 165. Honolulu.

Jannasch, H. W. 1967. Growth of marine bacteria at limiting concentrations

of organic carbon in seawater. Limnol. Oceanogr. 72:264-271.

__________. 1970. Threshold concentrations of carbon sources

limiting bacterial growth in seawater. In 0. W. Hood (Ed.) Organic Matter in Natural Waters. Inst. Mar. Sci. (Alaska). Occ. Publ. #1.

__________ and G. E. Jones. 1959. Bacterial populations in

seawater as determined by different methods of enumeration. Limnol. Oceanogr. 4:128-139.

Johnson, L. F. and E. A. Curl. 1972. Methods for Research on the

Ecology of Soil-Borne Plant Pathogens. Burgess Publishing Co. Minneapolis.

Jones, D. G. and P. E. Russel. 1972. The application of immunofluorescence

techniques to host plant/nodule bacteria selectivity experiments using Trifolium repens. Soil Biol. Biochem. 4:277282.

Josey, D. P., J. L. Beynon, A. W. B. Johnston and J. E. Beringer.

1979. Strain identification in Rhizobium using intrinsic antibiotic resistance. J. Appl. Bacteriol. 46:343-350.

Kishinevsky, B. and M. Bar-Joseph. 1978. Rhizobium strain identification in Arachis hypogea nodules by enzyme-linked immunosorbant assay (ELISA). Can J. Microbiol. 24:1537-1543.

Kleczkowski, A. and H. G. Thorton. 1944. A serological study of

root-nodule bacteria from pea and clover inoculation groups. J. Bacteriol. 48:661-672.

Koontz, F. P. and J. E. Faber. 1961. Somatic antigens of Rhizobium

japonicum. Soil Sci. 91:228-252. Lahav, N. 1962. Adsorption of sodium bentonite particles on

Bacillus subtilis. Plant Soil. 17:191-208. Lammana, C. and M. F. Mallette. 1965. Basic Bacteriology:

Its Biological and Chemical Background. The Williams & Wilkins Co. Baltimore.

Lehninger, A. L. 1970. Biochemistry: The Molecular Basis of

Cell Structure and Function. Worth Publishers Inc. New York. Leonard, L. T. 1943. A simple assembly for use in testing cultures

of rhizobia. J. Bacteriol. 45:523-525. Lindermann, W. C., E. L. Schmidt and G. E. Ham. 1974. Evidence for

double infection within soybean nodules. Soil Sci. 118:274-279. Litchfield, C. D., J. B. Rake, J. Zindulis, R. T. Watanabe and D. J. Stein.

1975. Optimization of procedures for recovery of heterotrophic bacteria from marine sediments. Microbiol. Ecol. 1:219-233.

Marshall, K. C. 1967. Electrophoretic properties of fast- and

slow-growing species of Rhizobium. Aust. J. Biol. Sci. 20: 429-438.

__________. 1968. Interaction between colloidal montmorillonite

and cells of Rhizobium species with different ionogenic surfaces. Biochim. Biophys. Acta. 156:179-186.

__________. 1969a. Orientation of clay particles sorbed on

bacteria possessing different ionogenic surfaces. Biochim. Biophys. Acta. 193:472-474.

Marshall, K. C. 1969b. Studies by microelectrophoretic and

microscopic techniques of the sorption of illite and montmorillonite to rhizobia. J. Gen Microbiol. 56:301-306.

__________. 1971. Sorptive interactions between soil particles

and microorganisms. In A. D. McLaren and J. Skujins (Eds.) Soil Biochemistry. Vol. 2. Marcel Dekker. New York. pp. 409- 445.

__________. 1976. Interfaces in Microbiol Ecology. Harvard

University Press. Cambridge.

__________. 1980. Adsorption of microorganisms to soils and sediments. In G. Bitton and K. C. Marshall (Eds.) Adsorption of Microorganisms to Surfaces. John Wiley & Sons. New York. pp. 317-329.

Martinez-Drets, G. and A. Arias. 1972. Enzymatic basis for

differentiation of Rhizobium into fast- and slow-growing groups. J. Bacteriol. 109:467-470.

__________, A. Gardiol and A. Arias. 1977. 6-Phospho-D-

Gluconate: NAD+ 2-Oxido-reductase (Decarboxylating) from slow- growing rhizobia. J. Bacteriol. 130:1139-1143.

Masterson, C. L. and M. T. Sherwood. 1974. Selection of Rhizobium

trifolii strains by white and subterranean clovers. In J. Agr. Res. 13:91-99.

May, S. N. 1979. Ecological studies on lentil rhizobia: Competition

and persistence in some tropical soils. Master’s Thesis. University of Hawaii. Honolulu.

Means, U. M. and H. W. Johnson. 1968. Thermostability of antigens

associated with serotype of Rhizobium japonicum. Appl. Microbiol. 16:203-206.

__________, __________ and R. A. Date. 1964. Quick serological

method for classifying strains of Rhizobium japonicum. J. Bacteriol. 87:547-553.

Medhane, N. S. and P. L. Patil. 1974. Comparative performance of

different isolates of gram (Cicer arietinum L.) Rhizobium in pot culture and field experiments. Pesticides. 8:48-49.

Metzler, D. E. 1977. Biochemistry: The Chemical Reactions of

Living Cells. Academic Press. New York. Michaels, A. S. 1958. Deflocculation of kaolinite by alkali

polyphosphates. Ind. Eng Chem. 50:951-958. Niepold, F., R. Conrad and H. G. Schlegel. 1979. Evaluation of the

efficiency of extraction for the quantitative estimation of hydrogen bacteria in soil. Antonie van Leewenhoek. 45:485-497.

Pankhurst, C. E. 1979. Some antigenic properties of cultured cell

and bacteroid forms of fast- and slow-growing strains of Lotus rhizobia. Microbios. 24:19-28.

Pattison, A. C. and F. A. Skinner. 1973. The effects of antimicrobiol

substances on Rhizobium spp. and their use in selective media. J. Appl. Bacteriol. 37:239-250.

Pearl, H. W. 1975. Microbiol attachment of particles in marine and

fresh water ecosystems. Microbiol. Ecol. 2:73-83. Perfil’ev, B. V. and D. R. Gabe. 1969. Capillary methods of

investigating microorganisms. University of Toronto Press. Toronto.

Pramer, D. and E. L. Schmidt. 1964. Experimental Soil Microbiology. Burgess Publishing Co. Minneapolis.

Reed, W. M. and P. R. Dugan. 1978. Distribution of Methylomonas

methanica and Methylosinus trichosporium in Cleveland harbor as determined by an indirect fluorescent antibody-membrane filter technique. Appl. Environ. Microbiol. 35:422-430.

Reyes, V. G. and E. L. Schmidt. 1979. Population densities of

Rhizobium japonicum strain 123 estimated directly in soil and rhizospheres. Appl. Environ. Microbiol. 37:854-858.

Rubertschik, L., M. B. Roisin and F. M. Bieljansky. 1936. Adsorption

of bacteria in salt lakes. J. Bacteriol. 32:11-31. Sanchez, P. A. 1976. Properties and Management of Soils in the

Tropics. John Wiley & Sons. New York. Scheraga, M., M. Meskill and C. D. Litchfield. 1979. Analysis of

methods for the quantitative recovery of bacteria sorbed onto marine sediments. In C. J. Litchfield and P. L. Seyfried (Eds.) Methodology of Biomass Determinations and Microbial Activities in Sediments. American Society for Testing and Materials. Philadelphia. pp. 21-39.

Schmidt, E. L. 1973. Fluorescent antibody techniques for the study

of microbiol ecology. Bull. Ecol. Res. Comm. (Stockholm). 19:67-76.

__________. 1974. Quantitative autecological study of microorganisms

in soil by immunofluorescence. Soil Sci. 118:141-149. __________. 1978. Ecology of the legume root-nodule bacteria.

In Y. R. Domergues and S. V. Krupa (Eds.), Interactions Between Non-pathogenic Soil Microorganisms. Elsevier. Amsterdam. pp. 269-303.

__________. 1979. Initiation of plant root-microbe interactions.

Ann. Rev. Microbiol. 33:355-376. Schmidt, E. L., R. 0. Bankole and B. B. Bohlool. 1968. Fluorescent antibody

approach to study rhizobia in soil. J. Bacteriol. 95:1987-1992.

Sherwood, M. T. and C. L. Masterson. 1974. Importance of using

the correct test host in assessing the effectiveness of indigenous populations of Rhizobium trifolii. Ir. J. Agr. Res. 13:101-108.

Skinner, F. A., P. C. T. Jones and J. E. Mollison. 1952. A

comparison of a direct- and plate-counting technical for the quantitative estimation of soil microorganisms. J. Gen. Microbiol. 6:261-271.

Skrdleta, V. 1969. Serological analysis of eleven strains of

Rhizobium japonicum. Antonie van Leewenhoek. 35:77-83.

Soil Conservation Service. 1972. Soil Survey of Islands of Kauai, Oahu, Maui, Molokai and Lanai, State of Hawaii. Soil Conservation Service. United States Department of Agriculture. Washington, D.C.

Soil Conservation Service. 1976. Soil Survey and Laboratory Data

for Some Soils of Hawaii. Soil Conservation Service. United States Department of Agriculture. Washington, D.C.

Stevens, J. W. 1923. Can all strains of a specific organism be

recognized by agglutination? J. Infectious Diseases. 33:557- 566.

Stotzky, G. 1966a. Influence of clay minerals on microorganisms.

II. Effect of various clay species, homoionic clays, and other particles on bacteria. Can. J. Microbiol. 12:831-848,

__________. 1966b. Influence of clay minerals on microorganisms.

III. Effect of particle size, cation exchange capacity and surface area on bacteria. Can. J. Microbiol. 12:1235-1246.

__________. 1972. Activity, ecology and population dynamics of

microbes in soil. Crit. Rev Microbiol. 2:59-137. __________. 1973. Techniques to study interactions between

microorganisms and clay minerals in vivo and in vitro. Bull. Ecol. Res. Comm. (Stockholm). 17:17-28.

__________ and L. T. Rem. 1966. Influence of clay minerals on microorganisms.

I. Montmorillonite and kaolinite on bacteria. Can. J. Microbiol. 12:547-563.

Trinick, M. J. 1969. Identification of legume nodule bacteroids

by the fluorescent antibody reaction. J. Appl. Bacteriol. 32:181-186.

van Brussel, A. A. N., K. Planque, and A. Quispel. 1977. The wall

of Rhizobium leguminosarum in bacteriod and free-living forms. J. Gen. Microbiol. 101:51-56.

van der Maesen, L. J. G. 1972. Cicer L., a monograph of the genus,

with special reference to the chickpea (Cicer arietinum L.) its ecology and cultivation. Mededelingen Landbouwhogeschool Wageningen, Nederland, Publication No. 72-10. Veenmen and Zonen. Wageningen.

Vidor, C. and R. H. Miller. 1979a. Efficiency of the quantitative

fluorescent antibody technique (FA) for ecologic studies of Rhizobium japonicum. Soil Biol. and Biochem. In Press.

____________________. 1979b. Relative saprophytic competence

of Rhizobium japonicum strains in soils as determined by the quantitative fluorescent antibody technique (FA). Soil Biol. Biochem. In Press.

____________________. 1979c. Influence of the soybean host on the population of Rhizobium japonicum in soil. Soil Biol. Biochem. In Press.

Vincent, J. M. 1962. Influence of calcium and magnesium on the

growth of Rhizobium. J. Gen. Microbiol. 28:653-663. __________. 1970. A Manual for the Practical Study of Root-Nodule

Bacteria. IBP Handbook 15. Blackwell Scientific Publications. Oxford.

__________. 1977. Rhizobium: General microbiology. In

R. W. F. Hardy and W. S. Silver (Eds.). A Treatise on Dinitrogen Fixation. Section III. John Wiley & Sons. New York.

__________ and B. A. Humphrey. 1968. Modification of the antigenic

surface of Rhizobium trifolii by a deficiency of calcium. J. Gen. Microbiol. 54:397-405.

__________ and __________. 1970. Taxonomically significant

group antigens in Rhizobium. J. Gen Microbiol. 63:379-382. __________, __________ and V. Skrdleta. 1973. Group

antigens in slow-growing rhizobia. Arch. Microbiol. 89:466-469. Wollum, A. G. II. and R. H. Miller. 1980. A density centrifugation

method for recovering Rhizobium spp. from soil for fluorescent antibody studies. Appl. Environ. Microbiol. 39:466-469.

Wright, W. H. 1925. The nodule bacteria of soybeans: I.

Bacteriology of strains. Soil Sci. 20:95-141. ZoBell, C. E. 1943. The effect of solid surfaces upon bacterial

activity. J. Bacteriol. 46:39-56.